Skip to main content

Full text of "Journal of the Chemical Society"

See other formats

This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project 
to make the world's books discoverable online. 

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject 
to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books 
are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover. 

Marks, notations and other marginalia present in the original volume will appear in this file - a reminder of this book's long journey from the 
publisher to a library and finally to you. 

Usage guidelines 

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the 
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we have taken steps to 
prevent abuse by commercial parties, including placing technical restrictions on automated querying. 

We also ask that you: 

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for 
personal, non-commercial purposes. 

+ Refrain from automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine 
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the 
use of public domain materials for these purposes and may be able to help. 

+ Maintain attribution The Google "watermark" you see on each file is essential for informing people about this project and helping them find 
additional materials through Google Book Search. Please do not remove it. 

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just 
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other 
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of 
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner 
anywhere in the world. Copyright infringement liability can be quite severe. 

About Google Book Search 

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers 
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web 

at http : //books . google . com/| 

f . www lit 

wr %t • r » rwt \y I %f ^ ^ I %f ''' 

rrw «^irir 

I Milligram Molecule in 100 

in SCO " 

Journal of the Chemical Society 


)val bocietv 





-^ t- ■*— ^ 


.^j ;. y^ro-i 


Digitized by ' 




Digitized by VjOOQIC 

Digitized by VjOOQIC 



/ ' I 


HoKACK T. B»owH, LL.D., F.R.S. 
J. I>BWAR, L.L.I>., F.R.S. 
Wtndham R- Ditnstan, M.A., F.R.S. 
C. E- Grovks, F.B-S. 
A. Vmtifoii HiLKCOintT, M.A., F.R.S. 
C. T. Mexcjock, M.A., F.E.S. 

%vtmi!i\u sA ^nbluHtum; 

R. Meldola, F.R.S. 


A. Scott, D.Sc, F.RS. 

T. E. Thorpe, LL.D., F.R.S. 

W. A. TiLDEN, D.Sc. F.R.S. 

C. K Groves, F.R.S. 

W. P. Wynne, D.Sc, F.B.S. 

A. J. Grssnaway. 

1899. Vol. LXXV. 



Digitized by VjOOQIC 


Digitized by VjOOQIC 





tmtmaittt oi ^nblitistvmi 

Ho&ACS T. BaowN, LL.D., P.R.S. 
J. TfmwASL, I.L..I>., F.R.S. 
yrmmiKAM. 'EL DTTKBTAir, M.A., F.ItS. 
C. E. Gkovbs, F.B.S. 
A. Ymkjsov Harcjoitbt, M.A., F.R.S. 
C. T. HLbtooox, M.A., F.K.8. 

R. Mbldola, F.R.S. 


A. Scott, D.Sc, F.B.S. 

T. E. Thobpb, LL.D., F.E.S. 

W. A. TiLDBN, D.Sc.-, F.R.S. 

C. E. Geotes, F.R.S. 

W. P. WTimB, D.Sc, F.R.8. 

A. J. Gbbenawat. 

Vol. LXXV. Parti. 



Digitized by VjOOQIC 

UiOHARD Cult and Buns, Liuitbd, 


Digitized by VjOOQIC 





tsrmaaiUt of ^ttblitatton ; 

Mo&ACK T. Browk, LL.D., P.E.S. 
J. Okwar, LLu D., F.R.8. 
TV^-raDHiLic R. DiTNSTAir, H.A., F.R.S. 
C. K. Grovks, F.R.S. 
A. Vkrnon^ Harcoitet, M.A., F.B.S. 
C T. Heycock, M.A., F.R.S. 

R. Meldola, F,R.S. 


A. SooTT, D.So., F.B.S. 

T. E. Thorpe, LL.D., F.B.S. 

W. A. TiLDBN, D.Sc., F.R.S. 

Cbtlors : 

C. S. Oroybs, F.R.S. 

W. P. Wynne, D.Sc, F.R.S. 



1899. Vol. LXXV. Part II. 



Digitized by VjOOQIC 

JliOHABO Clay and Som, Lxmitbd 


Digitized by VjOOQIC 




I. — ^The Oxidation of Polyhydric Alcohols in Presence of Iron. 

By Hbkrt J. HoBSTMAK Fbkton, M.A.^ and Henbt 

Jackson, B.A., B.Sc. 1 

n.— j8-Aldebydopropionic Acid, CHO-CHg-CHj-COOH, and 

/8-Aldehydoisobutyric Acid, CHO-CH2-CH(CH3)-COOH. 

By William H. Pebkin, jun., and Chablbs H. G. SpbaiYklimg 1 1 
IIL — ^Gannabinol. Part I. By Thomas Bablow Wood, M.A. ; 

Wi. T. Newton Spiybt, M.A. ; Thomas Hill Eastbbfield, 

M.A., Ph.D 20 

lY. — Characterisation of Racemic Compounds. By Fbedebio 

Staklet Kipping and William Jaokson Pope ... 36 
Y. — Crystalline Form of Iodoform. By William Jackson 

Pope 46 

YI. — ^j8^-Dimethylglataric Acid and its Derivatives ; Synthesis 

of cU- and Iratw-Caronic Acids. By William H. Pebkin, 

Jan., and Joceltn IF. Thobpe 48 

' Yn. — Synthesis of a)9)8-Trimethylglutaric Acid, 

I C00H-CH(CHg)-C(CH,)2-CH,-C00H. 

\ By William H. Pebkin, jun., and Jocelyn F. Thobpe . 61 

YIII. — Occurrence of Orthohydroxyacetophenone in the Volatile 
r Oil of CMane glabra. By Wyndham R. Dunstan, F.RS., 

I I and Thomas Andebson Hbnby, Salters' Research Fellow 

1' in the Laboratories of the Imperial Institute ... 66 

! IX. — Occurrence of Hyoscyamine in the Hyo9cyamu$ mulietu 
^i of India. By Wyndham R. Dunstan, F.R.S., and Habold 

I Bbown, Assistant Chemist in the Laboratories of the 

', Imperial Institute . 72 

. I X. — Preparation of Hyponitrite from Nitrite, through Hydroxy- 

amidosulphonate. By Edwabd Divebs, M.D., D.Sc., F.RS., 
and Tamemasa Haga, D.Sc., F.C.S. . . . . .77 

XL — ^Absorption of Nitric Oxide in Gas Analysis. By Edwabd 

,^ DiYEBS, M.D., D.Sc., F.RS 82 

'I XIL — Interaction of Nitric Oxide with Silver Nitrate. By 

Edwabd Divebs, M.D., D.Sc., F.RS 83 

I A 2 

Digitized by VjOOQIC 


XIII. — ^Preparation of Pure Alkali Nitrites. By Edward 

DivBES, M.D., D.Sc., F.R.S 85 

XrV". — Reduction of an Alkali Nitrite by an Alkali Metal. By 

Edward Divers, M.D., D.Sc, F.R.S 87 

XV, — Hyponitrites ; their Properties, and their Preparation 
by Sodium or Potassium. By Edward Divers, M.D., 
D.Sc., F.R.S. 95 

XVI. — ^Derivatives of Camphoric Acid. Part III. By Fred- 
eric Stanley Kipping, Ph.D., D.Sc., F.R.S. . . .125 

XVII. — a-Ketotetrahydronaphthalene. By Frederic Stanley 

Kipping, Ph.D., D.Sc, F.R.S., and Alfred Hill . . 144 

XVni. — Production of Optically Active Mono- and Di-alkyl- 
/ ozysuccinic Acids from Malic and Tartaric Acids. By 
Thomas Purdie, F.R.S., and William Pitkeathly, B.Sc, 
Berry Scholar in Science 153 

XIX. — ^Determination of the Constitution of Fatty Acids. 
Part I. By Arthur William Crossley and Henry 
Rondel Le Sueur 161 

XX. — Some Halogen Derivatives of Acetone Dicarboxylio 

Acid. Part I. By Frederick W. Dootson, M.A. . . 169 

XXI. — Action of Chlorosulphonic Acid on the Paraffins and 
other Hydrocarbons as a means of Purifying the Normal 
Paraffins. By Sydney Young, D.Sc, F.R.S. . . .172 

XXII. — Oxidation of Sulphocamphylic Acid. By W. H. Perkin, 

jun. 175 

XXIII. — ^Researches on Moorland Waters. I. Acidity. By 

William Ackroyd, F.I.C. 196 

XXIY.— The Nutrition of Yeast. Part I. By Arthur L. 

Stern, D.Sc 201 

XXV. — An Isomeride of Amarine. By H. Lloyd Snapb, D.Sc, 

Ph.D., and Arthur Brooke, Ph.D 208 

XXYI. — Studies of the Terpenes and Allied Compounds. 
Nitrocamphor and its Derivatives. lY. Nitrocamphor as 
an Example of Dynamic Isomerism. By T. Martin 
LowRY, B.Sc 211 

XXYII. — ^The Action of Ammonia on Ethereal Salts of Organic 

Acids. By Siegfried Ruhemann 245 

XXYin. — The Changes of Yolume due to Dilution of Aqueous 

Solutions. By Edward Bruce Herschbl Wade, M.A. . 254 

XXIX — Methanetrisulphonic Acid. By Ernest Harold 

Bagnall, B.Sc 278 

XXX. — Maltodextrin : its Oxidation-Products and Constitution. 

ByHoRACET.BROWN,LL.D.,F.R.S.,and Jambs Hills Millar 286 

Digitized by VjOOQIC 



XXXL — Attempts to Prepare Pure Starch Derivatives through 
their Nitrates. By Hobaob T. Brown, LL.D., E.II.S., and 
Jaus Htllb Millab 308 

XXXn. — The Stable Dextrin of Starch Transformations, and 
its Relation to the Maltodeztrins and Soluble Starch. By 
HoKACB T. Bbown, LL.D., F.B;S., and James Hills Millar 315 

XXXm. — Position Isomerism and Optical Activity; the 
Methylic and Ethylic Salts of Benzoyl- and of Ortho-, Meta-, 
and Para-toluylmalic Acids. By Percy Frankxaih), Ph.D., 
F.B.S./aiid Frederick Malcolm Wharton, A.I.C., late 
Priestley Scholar in Mason University College, Birmingham. 337 

XXXrV. — Some Begularities in the Botatory Power of Homo- 
logous Series of Optically Active Compounds. By Percy 
Frankland, Ph.D., F.R.S 347 

XXXV. — ^Detection and Determination of Sucrose in the 

Presence of Lactose. By Edwin Dowzard . . .371 

XXXVI. — Note on Certain. Isomeric Tertiary Benzylthioureas. 

By Augustus Edward Dixon, M.D 373 

XXXVIL — On Lossner's Benzoylethyloxysulphocarbamic Acid ; 
and the Formation of Pseudoureas. By Augustus Edward 
Dixon, M.D 375 

XXXVin. — Action of Metallic Thiocyanates on Certain Sub- 
stituted Carbamic and Ozamic Chlorides; and a New 
Method for the Production of Thiobiurets. By Augustus 
E. Dixon, M.D 388 

ICTCTCIX. — Formation of a-Pyrone Compounds and their Trans- 
formation into Pyridine Derivatives. By Siegfried 

XL. — Hydrolysis of the y-Cyanides of Di-substituted Aceto- 

acetates. By William Trevor Lawrence . . .417 

XLI. — Bromomethylfurfuraldehyde. By Henry J. Hobstman 
Fbnton, M.A., and Miss Mildred Gostling, B.Sc, Bathurst 
Student of Newnham College 423 

XLII. — ^A Reaction of some Phenolic Colouring Matters. By 

Arthur George Perkin, F.R.S.E 433 

XLUI. — ^A Method of Studying Polymorphism, and on Poly- 
morphism as the Cause of some Thermal Peculiarities of 
Chloral Hydrate. By William Jackson Pope . 455 

XLIV. — Contribution to the Characterisation of Bacemic 

Compounds. By A. Ladenburg 465 

XLT. — Etherifieation Constants of Substituted Acetic Acids. 

By John J. Sudborouoh and Lorenzo L. Llotd 467 

Digitized by VjOOQIC 

XLVI. — ^The Eotatory Powers of Optically Active Methoxy- 
and Ethozy-propionic Acids prepared from Active Lactic 
Acid. By Thomas Pubdie, F.R.S., and James C. Ibvinb, 6.Sc. 483 

XLYII. — Position-Isomerism and Optical Activity. The Com- 
parative Rotatory Powers of Methylic and Ethylic Ditoluyl- 
glycerates. By Percy Fbankland, F.R.S., and Henry 
Abton, late Priestley and Forster Scholar in Mason Uni- 
versity College, Birmingham 493 

XLVIII. — Isomeric Fencholenic. Acids. By George Bertram 

CocKBURN, B.A., B.Sc. 501 

XLIX. — Synthesis of some Derivatives of )9)8'-Dipyridyl from 
Citraziaic Acid. By W. J. Sell, M.A., F.I.O., and 
H. Jackson, B.A., B.Sc. ....... 507 

L. — ^The Condensation of Oxalic Acid and Resorcinol. By John 

Theodore Hewitt and Arthur Ernest Pitt . . , 5J6 

LI. — Synthesis and Preparation of Terebic and Terpenylic Acids. 

By W. Trevor Lawrence 527 

LII. — Ethyl Ammoniumsulphite. By Edward Divers and 

Masataka Ogawa 533 

LIII. — Ethyl Ammonium Selenite and the Non-existence of 
Amidoselenites (Selenosamates). By Edward Divers and 
Seihachi Hada. . 537 

LIV. — The Action of Certain Acidic Oxides on Salts of Hydroxy- 
acids. IV. By George Gerald Henderson, D.Sc., M. A., 
Thomas Workman Orr, and Robert J. Gibson Whitehead 542 

LY. — Derivatives of oa'-Dihromocamphorsnlphonic Acid. By 

Arthur Lapworth 558 

LVI. — Crystalline Glycollic Aldehyde. By Henry J. Horstman 

Fbnton, M.A., and Henrt Jackson, B. A., B.Sc. . . . 575 

LVII. — ^Diortho-substituted Benzoic Acids. P^urt IV. Form- 
ation of Salts from Diortho-substituted Benzoic Acids and 
different Organic Bases. By Lorenzo L. Llotd and John 
J. SudboeOugh 580 

L VIIL — A New Compound of Arsenic and Tellurium. By E. C. 

SzARVASY, Ph.D., and C. Messinger, Ph.D. . . . 597 

LIX. — The Combustion of Carbon Disulphide. By Harold 

Baily Dixon and Edward John Bussell .... 600 

LX. — ^The Action of Nitric Oxide on Nitrogen Peroxide. By 

Harold Baily Dixon and James Dysart Peterrin • .613 

LXI. — ^On the Mode of Burning of Carbon. By Harold Baily 

Dixon 630 

Digitized by VjOOQIC 



LXn. — A study of the Absorption Spectra of Isatin, Car- 
bostyril, and their Alkyl Derivatiyes iu Relation to Taat- 
omerism. By Walter Nosl Habtlet, F.RS., and James 
J. DoBBiE, M. A., D.Sc. 640 

LXnL — ^Preparation of Acid Phenylic Salts of Dibasic Acids. 

By Samuel Barkett Schbtybb, D.Sc., Ph.D. . .661 

LXIV.— Oorydaline. Part VI. By James J. Dobbib, D.Sc., 

M.A., and Alezakdeb Laubeb 670 

LXV. — ^The Relative Efficiency and UsefulDess of various 
Forma of Still-head for Fractional DistillatioD, with a 
Description of some new Forms possessing special Advan- 
tages. By Sydnet Young, D.Sc., F.KS 679 

LXVI. — The Salts of Dimethylpyrone, and the Quadrivalence 
of Oxygen. By J. N. Collie, Ph.D., F.R.S., and Thomas 
Tickle, Salters' Company's Research Fellow in the Research 
Laboratory of the Pharmaceutical Society of Great Britain 710 

LXYII. — Chemical Examination of the Oleo-resin of Dacryodeg 

hexandra. By Andbew Mobb, A.R.C.S 718 

LXYin. — ^Sstimation of Boric Acid mainly by Physical 

Proceases. By A. Wynteb Blyth 722 

LXIX.— The Blue Salt of Fehling's Solution and other Cupro- 
tartrates. By Obme Masson, M.A., D.Sc., and B. D. . 
Steels, B.Sc 725 

LXX. — ^The Allotropic Modifications of Phosphorus. By David 

Lbohabb Chapman, B.A. 734 

LXXL — Oxidation of Furfuraldehyde by Hydrogen Peroxide. 

By C. F. Cboss, E. J. Bevak, and Thv. Heibbbg . 747 

LXXIL — Active and Inactive Phenylalkyloxyacetio Acids. By 

Alex. McKenzie, M.A«, D.Sc 753 

LXXIIL — Some Derivatives of Dimethyldihydroresorcinol. By 

Abthub William Cbosslby . ... . 771 

LXXIY.— Condensation of Ethylic Salts of Acids of the 
Acetylene Series with Ketonic Compounds. By Sibqfbibd 
RuHEMABN and A« Y. Cunnikgton ..... 778 

LXXV. — Action of Hydrogen Peroxide on Carbohydrates in the 
Presence of Ferrous Salts. By Robebt Selby Mobbbll, 
M.A., Ph.D., and James Mubbat Cbofts, B.A., B.Sc. . 786 

LXXYI. — ^The Action of Alkyl Haloids on Hydroxylamine. 
Formation of Substituted Hydroxylamines and Oxamines. 
By Wtnbham R. Duustan, F.R.S., and Ebiobst Couldino, 
B.Sc 792 

LXXYII. — ^The Condeusation of Ethylic Acetonedicarboxylate 
and Constitution of Triethyllc O^cinoltricarboxylate. By 
David Smiles Jebdah 808 

Digitized by VjOOQIC 



LXXVIIL — ^The Colouring Matter of Cotton Flowers, Oouypwm 
herbaceum. Note on Bottlerin. B7 Abthub Obobge 
Pbekin, F.R.S.E 826 

LXXIX. — The Colouring Matters contained in Dyer's Broom 
(Geniita tinctoria) and Heather (CaUuna vulgarig). By 
Abthub Geobob Pbrkin, F.B.S.E., and Fbedbbice Geobge 
Nbwbubt 830 

LXXX. — Researches on the Alkyl-substituted Succinic Acids. 
Part I. Methods of Preparation. By William A. Bonb 
and Chables H. G. Spbakeling 839 

TjXXXL — Some Derivatives of Dibenzyl Ketone. By Fbancis 
E. Fbanois, B.Sc., Ph.D., Lecturer in Chemistry, University 
College, Bristol 866 

LXXXII. — Action of Light and of Oxygen on Dibenzyl Ketone. 

By Emily C. Fobtet, B.Sc 871 

LXXXIII. — ^The Vapour Pressures, Specific Volumes and 
Critical Constants of Hezamethylene. By Stdnbt Youno, 
D.Sc., F.R.S., and Emily C. Fobtby, B.Sc. . . .873 

LXXX IV. — ^The Composition and Tensions of Dissociation 
of the Ammoniacal Chlorides of Cadmium. By William 
Bobebt Lang, D.Sc, and Albbbt Rigaut .... 883 

TjXXXV. — The Aluminium-Mercury Couple. Part I. Action 
of Sulphur Chloride on some Hydrocarbons in presence of 
the Couple. By Julius Bbbbnd Cohen and Fbedbbick 
William Skibbow, The Yorkshire College .... 887 

LXXXVL— The Aluminium-Mercury Couple. Part II. The 
Action of Bromine on Organic Compounds in presence of 
the Couple. By Julius Bbbend Cohen and Hbnby D. 
Dakin, The Yorkshire College 893 

LXXXVII. — Experiments on the Constitution of Isocamphor- 
onic Acid. By Willlam Hbnby Pbbkin, jun., and Jocblyn 
Field Thobpb 897 

LXXXVIIL— The ow- and ^rowt-^-Phenylbutane-oaiOj-teicarb- 
oxylic Acids, 

By Jocblyn Field Thobpb and Willlam XJdall . . 904 

LXXXIX. — Experiments on the Synthesis of Camphoric Acid. 
Part II. By H. A. Audbn, William Hbnby Pebkin, jun., 
and J. L. Ross 909 

XC. — ^The Action of Ethylene Dibromide and Trimethylene Di- 
bromide on the Sodium Compound of Ethylic Cyanacetate. 
By H. C. H. Cabpentbb and William Hbnby Pebkin, jun. 921 

XCI. — Influence of Substitution on Specific Rotation in the 
Bomylamine Series. By Mabtin Onslow Fobsteb, Ph.D.y 
D..Sc. . . . 953 

Digitized by VjOOQIC 



XCn. — Studies of the Acids of the Acetylene Series. By 

Siegfried Buhehann and Alfred V. Cunnington. . 954 

XCIII. — A Contribution to the Chemistry of the Mandelic Acids. 

By Alex. McKenzie, M.A., D.Sc 964 

XCrV. — ^Non-existence of the so-called Suboxide of Phosphorus. 

By "David Leonard Chapman and F. Austin Lidbdry. 973 

XCV.— The Chlorine Derivatives of Pyridine. Part III. The 
Interaction of Chlorine and Pyridine Hydrochloride. By 
William James Sell, M. A., F.I.C., and Frederick William 
DooTsoN, M. A. . 979 

XUVl. — Homocamphoronic and Camphononic Acids. By 
Arthur Lapworth D.Sc., and Edgar M. Chapman, 
Burrough's Scholar in the Research Laboratory of the 
Pharmaceutical Society of Great Britain 986 

XCVII. — The Action of Hydrogen Peroxide on Secondary and 
Tertiary Aliphatic Amines. Formation of Alkylated 
Hydroxy lamines and Oxamines. By Wyndham R. Dunstan, 
F.R.S., and Ernest Gouldino, B.Sc ' . 1004 

XCVm. — ^Amidoamidines of the Naphthalene Series. By 

Raphael Meldola, F.R.S., and Percy Phillip Phillips. . 1011 

XCIX. — Condensations of Anhydracetonebenzil and its Ana- 
logues with Aldehydes. By Francis R. Japp, F.R.S., and 
Alexander Findlay, M. A., B.Sc 1017 

C. — ^Triphenyloxazolone. By Francis R. Japp, F.R.S., and 

Alexander Findlay, M.A., B.Sc 1027 

01. — Interaction of Phenanthraquinone, Acetopheuone, and 
Ammonia. By Francis R. Japp, F.R.S., and Andrew N. 
Meldrum, B.Sc 1032 

OIL — Furfuran Derivatives from Benzoin and Phenols. By 

Francis R. Japp, F.RS., and Andrew N. Meldrum, B.Sc. 1035 

OIXL — Interaction of Benzoin with Phenylenediamines. By 

Francis R. Japp, F.R.S., and Andrew N. Meldrum, B.Sc. 1043 

CIV. — A Series of Substituted Nitrogen Chlorides and their 
Relation to the Substitution of Halogen in Anilideu and 
Anilines. By F. D. Chattaway and K. J. P. Orton . . 1046 

CV- — Synthetical Preparation of Glucosides. By Huqh Ryan, 
M.A., 1851 Exhibition Scholar of the Queen's College, 
Galway 1054 

CVI. — ^The Action of Sulphuric Acid on Fenchone. By James 

E. Marsh 1058 

CVII. — On a Method for Providing a Current of Gaseous 
Chloroform mixed with Air in any desired proportion, and 
on Methods for Estimating the Gaseous Chloroform in the 
Mixtures. By A. Vernon Hahcourt, M.A., F.R.S., Lee's 
Reader in Chemistry at Christ Church, Oxford . . 1060 

Digitized by VjOOQIC 



CVIII. — ^The Application of Powerful Optically Active Acids to 
the Resolution of Externally Compensated Basic Substances. 
Resolution of Tetrahydroquinaldine. By William Jackson 
Pope and Stanley John Peachey 1066 

CDC. — The Application of Powerful Optically Active Acids to 
the Resolution of Externally Compensated Basic Substances. 
Resolution of Tetrahydroparatoluquinaldine. By William 
Jackson Pope and Edmund Milton Rich .... 1093 

ex. — The Application of Powerful Optically Active Acids to 
the Resolution of Externally Compensated Basic Substances. 
Resolution of Racemic Camphorozime. By William 
Jackson Pope 1105 

CXI. — Homogeneity of DextrolsBvo-a-phenethylamine Dextro- 
camphorsulphonate. By William Jackson Pope and 
Alpeed William BLabvey 1110 

CXII. — A Method for Discriminating between ** Non-racemic " 
and " Racemic " Liquids. By William Jackson Pope and 
Stanley John Peachey 1111 

CXIII. — The Characterisation of "Racemic" Liquids. By 

Fbedebic Stanley Kipping and William Jackson Pope . 1119 

CXIV. — Asymmetric Optically Active Nitrogen Compounds. 
Deztro- and Lffivo-a-benzylphenylallylmethylammonium 
Iodides and Bromides. By William Jackson Pope and 
Stanley John Peachey 1127 

CXY. — Tetrazoline. By Siegfeied Ruhemann and H. E. 

Stapleton, Scholar of St. John's College, Oxford . .1131 

CXVI. — Action of Hydrolytic Agents on a-Dibromocamphor. 
Constitution of Bromocamphorenic Acid. By Abthur 
Lapwoeth 1134 

CXVII. — Camphoroxime. Part III. Behaviour of Camphor- 
oxime towards Potassium Hypobromite. By Mabtin 
Onslow FoBSTEB, Ph.D., D.Sc. 1141 

CXYIII. — Influence of an Unsaturated Linking on the Optical 
Activity of Certain Derivatives of Bornylamine. By Mabtin 
Onslow FoBSTEE, Ph.D., D.Sc 1141 

CXIX. — The Interaction of Sodium Hydroxide and Benz- 
aldehyde. By Chables Alexandeb Kohn, Ph.D., B.Sc, and 
William Tbantom, Ph.D., B.Sc 1161 

CXX. — The Ultra-violet Absorption Spectrum of Proteids in 

Relation to Tyrosine. By A. Wynteb Blyth . . .116 

Annual General Meeting 116 

Digitized by VjOOQIC 





I. — The Oxidation of Polyhydric Alcohols in 
Presence of Iron. 

By Hkhbt J. HoBsniAN FBNTONy M.A., and Hsnbt Jacksok, 

B.A., B.So. 

Thb specific influence which ferrous iron exerts on the oxidation of 
certain hydroxy-acids has been pointed out by one of the authors in 
several previous communications (Trans., 1894, 65, 899 ; Proc., 1898, 
tc). In the case of tartaric acid, for example, two atoms of 
hydrogen are removed from a molecule of the acid with the pro* 
duction of dihydroxymaleic acid. For the purpose of bringing about 
this change, the presence of ferrous iron is absolutely essential, but 
its proportion bears little, if any, relation to the yield of the product, 
the influence of the iron being apparently of the nature usually 
described as ^ catalytic' The most efficient oxidising agent for the 
porpoee is found to be hydrogen dioxide, but a similar result 
obtained by chlorine, potassium permanganate, te., by electrolysis, 
and by atmospheric oxygen in presence of sunlight. 

The study of this oxidation process is now being extended to several 
other hydroxy-acids, and Messrs. Morrell and Cross have undertaken 
its application to certain carbohydrates. Aiming at a complete 
investigation of the reaction as regards various typical classes of 
hydroxy-compounds, the present authors are making a series of 
observations on the behaviour of various alcohols, and this com- 
munication deals with the results which have so far been obtained. 
It will be shown that, in the case of all the polyhydric alcohols 
examined, the presence of ferrops iron exerts a remarkable influence 

VOU LXXV. Digitized by€00gle 


on the ozidation by hydrogen dioxide, and that in its absence the 
results are practically nil. 

In carrying out the experiments, a weighed quantity of the alcohol 
was dissolved in water, mixed with a solution of ferrous sulphate 
(corresponding to about ^ ^ i atom Fe * per molecule of alcohol 
employed) and strong ('20 volume') hydrogen dioxide, previously 
standardised by permanganate, was gradually added ; in every case, a 
blank experiment was made with exactly similar proportions and 
under the same conditions, but omitting the iron. A blank experiment 
was also made, for comparison, with ferrous sulphate and hydrogen 
dioxide alone ; 5 grams of crystallised ferrous sulphate being dissolved 
in 20 c.a of water, and about 10 c.c. of the peroxide gradually added. 
The rise of temperature was very slight, and was quite insignificant 
in comparison with that obtained in the cases mentioned below ; the 
liquid gave only a brown colour or precipitate with Fehling's solution 
in the cold, without any sign of reduction, and gave no indication with 
Schiff's rosaniline test. 

Monhf/dric AleohoU, 

MethyUo, ethylio, propylic, isopropylic, and amylio alcohols were 
examined. The experiments were made by taking 10 grams of the 
alcohol, diluting with water, dividing the solution into two equal 
parts, to one of which 0*5 gram of crystallised ferrous sulphate, dis- 
solved in a little water, was added, and each part was then mixed 
with the standardised hydrogen dioxide in the proportion of 
1 mol. H3O2 to 1 mol. of the alcohol. No perceptible rise of 
temperature occurred in any case either in presence or absence 
of iron, and the solutions gave practically no indication with Fehling's 
solution or with SchifP's rosaniline test. The results were, in fact, 
entirely negative under the conditions of the experiments, at any rate, 
as regards the production of aldehydic substances, nor could acetone 
be detected in the case of isopropylic alcohol. 

Ethyhne Glycol 

On adding hydrogen dioxide to a strong aqueous solution of 
ethylene glycol, in presence of a ferrous salt, a very sensible rise of tem- 

* In the case of tartaric add, only ^^ atom Fe, or leas, waa employed fbr one 
molecule of acid, bat in the preaent inveatigation it was found advisable to use a 
larger proportion, and poasibly it may be advantageoua atill further to increase the 
quantity. The reason for thia difference is probably to be found in the &ct that, in 
tiie former caae, the ferric salt formed is almoat instantly reduced at the expense of 
a portion of the dihydrozymaleic acid (Proc., May, 1898, 119), a result which 
apparently is not produced by the oxidation products here mentioned. 

Digitized by VjOOQIC 


peratnre ooenrSy and heat continues to be evolved for a considerable 
time after the solutions are mixed ; the product almost immediately 
restores the colour to a rosaniline salt previously bleached by sulphur 
dioxide, and reduces Fehling's solution in the cold. Phenylhydrazine 
acetate, after a few minutes, gives a yellowish-brown, crystalline pre- 
ei^tate, which continues to separate at the ordinary temperature for 
some days. In an exactly parallel blank experiment, omitting the 
iron salt, there is no perceptible rise of temperature, and the solution 
does not reduce Fehling's solution in the cold ; it gives, however, a 
faint indication with the rosaniline test, and phenylhydrazine acetate 
produces after a time a small amount of precipitate somewhat 
resembling in appearance that above mentioned, but more disooloured 
and resinous. 

From these results, it would appear that, although glycol is very 
slightly oxidised by hydrogen dioxide alone, the action is insignificant 
in oomparison with that produced in presence of iron. 

In order to investigate the nature of the product of oxidation, 
6*2 grams of ethylene glycol were dissolved in 70 c.o. of water, and 
mixed with a solution of 6*2 grams of crystallised ferrous sulphate ; 
to the mixture was gradually added 65*6 cc. of hydrogen dioxide 
containing 0*0559 H^O^ per c.c. The temperature began to rise imme- 
diately, and increased considerably after a few minutes. After 
standing 1^ hours, a solution of sodium acetate was added (to remove 
free mineral acid present in the peroxide), and the solution was then 
mixed with 21*5 grams of phenylhydrazine dissolved in 50 per cent. 
acetic acid. The mixture was heated in a regulated water bath at 
39* for 5 hours, allowed to stand 12 hours at the ordinary tempera- 
ture, and the precipitate was then filtered off, washed, and dried in 
the air ; the weight of crude osazone thus obtained was 2*8 grams. 
A parallel experiment was made with a similar quantity of glycol, 
but omitting the ferrous salt, and using the same quantity of phenyl- 
hydrazine, Ac, and heating exactly as before ; the precipitate formed 
in this case weighed only 0*35 gram. 

The crude osazone obtained in the above manner was first rubbed 
with a very little cold absolute alcohol, which removed a little 
resxnons matter, and the undissolved yellow powder was twice recrys- 
tallised from hot absolute alcohol, from which it separated in yellow, 
transparent plates, melting sharply at 169*5^, and exactly resembling 
in every way the product previously obtained from glycollic aldehyde 
(Trana., 1895, €f7, 774). 

The substance, dried at 100°, gave the following results on analysis. 

0-0900 Bubetancegave 0*2326 CO^and OOiSOHsO ; C = 70*49 ; H = 5*92. 
Cj^Hj^N^ requires C = 70*58 ; H = 5*88 per cent. 

B 2 

Digitized by VjOOQIC 


The sabBtance is, therefore, glyozalofiazone, Ah'jjV.o H ' ^^^ 

the oxidation product from which it was obtained may consequently 
be either glyecUic aldehyde or glyoxal. . 

Another similar quantity of glycol was now oxidised in presence of 
iron as before, the mixture well shaken with pure chalk, filtered, and 
the filtrate evaporated to a syrup in a vacuum desiccator ; absolute 
alcohol and a few drops of ether were then added, and the precipitated 
calcium salts filtered off. The solution, on again evaporating in a 
vacuum, left a syrup soluble in alcohol but nearly insoluble in ether, 
and showing all the reactions of glycollic aldehyde ; it was partially 
volatile with steam, reduced Fehling's solution and ammoniacal silver 
nitrate in the cold, and immediately answered Schiff's rosaniline test. 

On shaking a strong aqueous solution of this syrup with excess of 
freshly prepared sodium hydrogen sulphite, no crystalline substance 
was obtained even on allowing the mixture to stand for some days. 
It was thought that, on using twice the proportion of hydrogen 
dioxide (2 mols.), glyoxal might perhaps be formed, but on making 
another similar experiment with this proportion, no compound with 
the sulphite could be obtained. The result of the oxidation conducted 
in the above way is, therefore, glycollic (Udchyde, 


When a solution of glycerol (1 mol.) is mixed gradually with 
strong (20 volume) hydrogen dioxide (1 mol.) at the ordinary tempera* 
ture, little, if any, change takes place. No alteration of temperature 
occurs, and the mixture, even after some days, gives very strong 
reactions of hydrogen dioxide ; Fehling's solution, in the cold, is not 
reduced, and the addition of phenylhydrazine acetate merely gives rise 
to an evolution of gas owing to the presence of hydrogen dioxide, and 
there is a slight precipitation of resinous matter. When, however, an 
exactly parallel experiment is made with the previous addition of a 
small quantity of ferrous sulphate in solution, a very powerful action 
results. The peroxide is now consumed almost immediately, and the 
temperature rises very considerably, in fact, if the peroxide is added 
quickly, and the solutions are strong, the vessel becomes almost too 
hot to be held in the hand. The solution, when cold, now strongly 
reduces Fehling's solution in the cold, immediately answers SchifiE's 
rosaniline test, and if it is mixed ^vith phenylhydrazine acetate, a 
bulky, yellowish-brown, crystalline precipitate begins to separate at 
the ordinary temperature, and continues to do so for several days. 

If a ferric salt be substituted for ferrous in the above experiment, 
the result is entirely negative. In order to investigate the nature of 

Digitized by VjOOQIC 



the prodoet, 10 grams of glycerol, diluted with an equal bulk of water, I 

was mixed with a solution of I gram of crystallised ferrous sulphate, \ 

and 45 c.c. of hydrogen dioxide (of the before-mentioned strength) 
was slowly added, the rise of temperature being moderated by 
immersing the Teasel in cold water. Sodium acetate and 20 grams of 
phenylhydrazine, as acetate^ were then added as before, and the 
mixture was left for some days at the ordinary temperature ; the pre- 
cipitate formed was then wadied, drainedi and dried in the air. It 
weighed 6-2 grams. After crystallising from 50 per cent, alcohol, 
and then twice from benzene, the substance was obtained in beautiful, 
golden, transparent prisms which melted at 130 — 131^. The sub* 
stance, dried at 100% gave the following results on analysis. 

L 01185 gave 0-2920 CO, and 0-0630 H^O. 

n. 0-0760 „ 14-5c.c.nitrogenat22°and764mm. C = 67-22;H = 6-90;N-2116. 
CijH^^N^O requires 0«6716 ; H-6-97 ; N« 20*89 per cent. 

This product, therefore, exactly coincides with glyeeroscuane, 
OH-CHj-CINaH-O^Hg y^^ ^^ obtained by Fischer and Tafel 

from < glyceroee ' {Ber., 1887, 20, 1088, and 1888, 21, 2634), by Piloty 
and Ruff from dihydroxyacetoxime (.6^., 1897, 30, 1662), and by 
Piloty from dihydroxyacetone (Ber., 1898, 81, 3165). 

The oxidation-product here obtained may, therefore, be dihydroxy- 
acetone, C0<^^;^5, glyceraldehyde, OH-CH<^|[^^, or, like 

Fischer's * glyceroee,' a mixture of these two. 

'Olyoeroee,' obtained by the interaction of lead glyceroxide and 
bromine, was shown to consist for the greater part of dihydroxy- 
acetone ; this was proved by the large yield of trihydroxyisobutyric 
add obtained by the hydrogen cyanide reaction {Ber., 1889, 22, 110) 
and was later confirmed by Piloty {loe. eii., 3163) from the yield of 
dihydroxyacetoxime. The presence of glyceraldehyde was probable 
bom the fact that» after separation of the trihydroxybutyric acid as 
eakiam salt, lead acetate gave, in the mother liquor, a flocculent pre- 
cipitate insc^uble in hot water and in dilute acetic add, differing in 
these respects from the trihydroxyisobutyrate. The quantity obtained 
by Fischer and Tafel was not suffident for comparison with ' erythro- 
glndc add,' and, moreover, the properties of this latter acid are not 
at all characteristie. 

A sharp distinction, however, between trihydroxyisobutyric acid and 
eijthrogludc acid exists in the fact that the neutral salts of the 
former are not precipitated by normal lead acetate. 

The product obtained in the present instance was further examined 
b the following way. About 50 grams of glycerol were oxidised in 

Digitized by VjOOQIC 


presence of iron in the manner described above, and the mixture, when 
cold, was well stirred with excess of barium carbonate ; the solution, 
after filtration, was evaporated to a small bulk by distillation under 
greatly diminished pressure at 50^, concentrated to a syrup in a vacuum 
desiccator, and then mixed with several times its volume of absolute 
alcohol and a little ether. After being allowed to stand, it was filtered 
from barium and iron salts, the filtrate again allowed to evaporate in 
a vacuum desiccator, and the residue, dissolved in a little absolute 
alcohol, was mixed with about three times its volume of anhydrous 
ether ; the precipitate was then allowed to settle, and the ether- 
alcohol solution evaporated in a vacuum desiccator as before. 
No signs of crystallisation could be observed, however, even after 
repeating the alcohol-ether treatment and allowing the syrup to remain 
for several days in the vacuum, neither oould crystallisation be brought 
about by cooling or by stirring with a drop of water. Piloty {Ber., 
1898, 31, 3164) obtained crystallised dihydroxyacetone (from dihy- 
droxyacetoxime) ; and Wohl {ibid,, 2394), from the acetal of glycer- 
aldehyde, obtained crystals corresponding in composition to the 
aldehyde. Wohl's crystalline compound, however, was shown by the 
cryoscopic method to have a ^ double ' formula, becoming ' single ' 
when kept for some time in aqueous solution. 

In the present <!ase, therefore, it was not certain that crystallisation 
should be expected if the substance is glyceraldehyde, but the absence 
of crystallisation may very probably be due to tiie presence of 
impurities such as traces of barium or glycerol, which, by the above 
treatment, might not be wholly removed. In order to attack the 
problem, therefore, the method employed by Fischer and Tafel was 
followed j for this purpose, the purified syrup obtained in the manner 
above described was dissolved in 100 c.c. of water, and the reducing 
power of the solution estimated by Fehling's solution* The quantity 
available corresponded to 6*2 grams of glucose ; this was mixed with 7 
grams of anhydrous hydrogen cyanide and heated in a stoppered bottle, 
in aregulated bath, at 60°for 12 hourB,and then for 12 hours more at 60**. 
The light brown solution was then evaporated somewhat in a vacuum to 
remove free hydrogen cyanide, the mixture placed in a freezing mixture, 
saturated with dry hydrogen chloride, and allowed to remain for 
12 hours at 0% and for 2 days more at the ordinary temperature; 
it was then evaporated as completely as possible on a water bath, the 
residue dissolved in about 200 c.c. of water, 45 grams of barium 
hydroxide, dissolved in hot water, added, and the mixture heated on a 
water batL Alter a few minutes, a pale yellow, fiocculent precipi- 
tate separated which was collected while hot, washed, and treated with 
a slight excess of hot dilute sulphuric acid ; the filtered solution was very 
cautiously treated with baryta-water to remove the excess of sulphuric 

Digitized by VjOOQIC 


add, filtered, md the filtrate shaken with pare precipitated calciom car- 
bonate. The clear liquid thus obtained was concentrated and allowed 
to rtand for 4 days, but there were no signs of crystallisationi so that 
trihydioxybatyric acid appears to be either absent, or present in small 
q[aAntity only. The solution of calcium salt was then decolorised 
with a little animal charcoal, and a solution of normal lead acetate 
added ; a white, flooculent precipitate was produced which appeared to 
be insDluble in hot water or in acetic acid. This precipitate was well 
washed with hoc water and dried at 160^. 

L 0-2693 gave 02303 PbSO^. Pb = 6067 per cent, 
n. 01190 „ 01060 PbSO^. Pb = 60-82 

Lamparter {Amudm^ 184, 261) obtained erythroglucic acid by oxida- 
tion of erythritol with nitric acid, and, by precipitation of the free 
acid by excess of lead acetate, prepared the lead salt which, when 
dried at 160°, gave Pb » 61 '3 per cent. The salt would appear, there- 
fore, to have the composition G^Hi^PbOg (the lead replacing in a 
hydroiyl as well as the carbozyl group), for which theory requires 
Pb=60-70 per cent. 

Another portion of the ether-alcohol solution, prepared as before, 
was mixed with the calculated quantity of hydrozylamine in alcoholic 
sdiition, distilled to a small bulk under greatly reduced pressure, and 
kept in a vacuum desiccator ; the syrup which was left showed praotic- 
ally no signs of crystallisation, even after long standing, stirring, dEc. 
From the above results, it appears probable that the product of oxi- 
dation here obtained is chiefly, if not entirely, glyceraldehyde. 


Experiments with this substance were made in exactly the same 
manner as in the previous cases, and here again it was found that 
practically no actiou occurs in absence of ferrous iron ; if present, 
however, there is energetic oxidation with considerable evolution of 
heat. The product strongly rednces Fehling's solution on warming, 
and restoTOB the colour to a rosaniline salt bleached by sulphur 
dioxide. Phenylhydraadne acetate, after a short time, gives a beautiful, 
golden, flooculent precipitate, which increases in quantity on heating. 

Ten grams of erythritol was dissolved in 300 cc. of water, mixed with 
5 grams of crystallised ferrous sulphate, and 50 cc. of hydrogen dioxide 
abwly added as before ; to the mixture, 25 grams of phenylhydraxin 
as aeetate was then added, and the whole allowed to remain for 24 hours 
at the ordinary temperature. The precipitate, after being collected, 
washed, drained, and dried in. the air, weighed 3*7 grams. The mother 
Uqaor was then heated on a water bath for \\ hours, and the precipitate 

Digitized by VjOOQIC 


collected, washed, and dried as before ; it weighed 3*3 grams, so that 
the total yield was aboat 7 grams. 

For analysis, the first product was selected, since that obtained by 
heating the mother liqaor was somewhat discoloured and looked leas 
pura It was first heated with a large volume of water, in which it 
slowly dissolved, and the filtered solution deposited a golden, floccolent 
precipitate on cooling ; this was collected, air-dried, and twice re- 
crystallised from hot benzene, from which it separated in golden, 
microscopic needles. 

After being dried at 100°, it melted sharply at 167°, and gave the 
following numbers on analysis. 

00920 gave 0-2178 00^ and 0-0496 H,0. C = 64-56 ; H = 6-91. 
Erythrosasone, Oj^H^gN^Oj, requires = 64*43 ; H = 6*04 per cent. 

From the good yield of osazone obtained in this experiment, the 
authors are encouraged to attempt the direct isolation of the oxidation 
product, and experiments will shortly be made in this direction. 


In this case, again, the presence of ferrous iron exerts a remarkable 
influence in bringing about the oxidation by hydrogen dioxide ; a very 
considerable rise of temperature occurs, and the mixture, when cold 
gives an almost immediate precipitate with phenylhydrazine, and after 
separation of the iron, strongly reduces Fehling's solution on warming. 
In absence of ferrous iron, no such changes take place. 

Twenty grams of pure recrystallised mannitol were dissolved in 
100 c.c. of water, mixed with 5 grams of crystallised ferrous sulphate 
and 60 c.c. of hydrogen dioxide in the same manner as before. The 
solution, after cooling, was made just alkaline with sodium carbonate, 
then acidified with acetic acid, and to the mixture 6 grams of 
phenylhydrazine as acetate were added. The separation of a yellow 
precipitate began after about 1 minute, and was allowed to continue 
for 1 hour ; it was then collected, drained with the aid of the pump, 
washed, and allowed to dry in the air. The crude, orange-coloured 
substance, which weighed 12-5 grams, was purified by triturating it 
with a little acetone, draining with the pump, and crystallising from 
boiling water ; the filtered solution, on cooling, deposited a yellowish- 
white, crystalline precipitate, which was recrystallised from hot 60 per 
cent, alcohol, and twice again from hot water with the addition of a 
little animal charcoal. In this way, the product was obtained in the 
form of almost white, crystalline plates, which turn yellow at about 
195° and melt fairly sharply at 1 97 — 198°. For analysis, the substance 
was dried at 100°. 

Digitized by VjOOQIC 


L 01206 gave 0-2346 COj and 00710 H^O. = 63-29 ; H = 6-67. 
II. 0-1020 gave 93 c.c. nitrogen at 17° and 766 mm. N= 10*66. 
Maonoae hydrazone, C^H^gN^Os, requires C » 63-33 ; H = 6*66 ; N » 10-39 per cent. 

Direct Preparation of Mannoee from [Mannitol. — As the yield of 
hydiazone obtained in the manner just described appears to be 
considerably superior to that which results from the oxidation of 
mannitol by nitric acid, it became a matter of interest to ascertain 
whether the sugar could be prepared directly from the mixture 
obtained from oxidation, without previously isolating the hydrazone. 
Fifty grams of mannitol were dissolved in a small quantity of 
water, mixed with a solution of ferrous sulphate, and 160 c.c. of 
hydrogen dioxide added as previously described. The mixture was 
then treated with excess of freshly precipitated barium carbonate, 
filtered, the clear solution evaporated to a small bulk at about 60^ under 
a pressure of about 30 mm., and the concentration then continued 
in a vacuum desiccator at the ordinary temperature until a thick 
syrup remained. This was treated with about ten times its volume of 
absolute alcohol, filtered, and an excess of anhydrous ether added to 
the clear filtrate ; this caused the separation of a white, flocculent 
mass, which, on standing, collected to a light yellow syrup. After 
pouring off the liquid, the syrup was dissolved in a very small quantity 
of water, the solution exposed in a vacuum desiccator, and the syrupy 
residue thus obtained was rubbed with absolute alcohol and ether, 
when it was converted into a pasty mass, which became a greyish- 
whitOy amorphous gum when kept in a vacuum over sulphuric acid. 
0*36 gram of this product was dissolved in a little water, mixed with 
0-6 gram of phenylhydrazine as acetate, and the mixture allowed to 
stand for one hour ; a little absolute alcohol was then added and the 
precipitate collected on a weighed filter and dried at 100°. The weight 
of hydraaone obtained was 0-66 gram. 

Wt. of substance : wt. of hydrazone obtained = 1 : 1*66. 
Theory requiring, mannose : hydrazone » 1 : 1*63. 

The result, in fact, appears to be mannose in a pure, or nearly pure, 


Sxaetly similar phenomena are observed when this substance is 
oxidised by hydrogen dioxide in presence of ferrous iron, the blank 
test giving negative results. Ten grams of duldtol dissolved in 
SOO CO. of water, was mixed with 6 grams of crystallised ferrous 
mdphate and 40 c.o. of hydrogen dioxide added. After cooling, 20 
grams of phenylhydrazine as acetate was added, the mixture heated on 

Digitized by VjOOQIC 


a water bath for one hour, and the yellowish^ flocculent precipitate thus 
formed, when collected, washed, and air-dried, weighed 3-8 grams. 

The crude osazone was rubbed with a little cold absolute alcohol, to 
remove resinous matter, and then twice recrystallised from hot 
absolute alcohol ; the beautiful yellow leaflets thus obtained were dried 
a 100° and analysed. 

0-1030 gave 0-2290 CO, and 00684 H^O. C « 60-63 ; H » 6*29. 
CigHj^Np^ requires 0*60-33 ; H»6'15 per cent. 

The substance, when quickly heated, melted sharply at 206°, and 
appears, therefore, to be identical with that which Fischer and Tafel 
obtained by oxidising duldtol with bromine {Ber., 1887, 20, 3390). 


Only a small quantity of this substance was available, but the results 
obtained were exactly analogous to those previously recorded. In 
presence of iron, heat was evolved and the solution then strongly 
reacted with phenylhydrazine, the blank experiment without iron 
showing practically no result. 

0*7 gram of sorbitol was dissolved in water, mixed with a solution 
of 0-3 gram of ferrous sulphate, and 10 c.c. of hydrogen dioxide 
added ; on adding phenylhydrazine acetate and heating the mixture 
on a water bath for one and a half hours, an osazone was obtained, 
which, after being washed, and recrystallised from ethylic acetate, 
melted sharply at 203°, and closely resembles glucosazone. 

On analysis, 

0-1190 gave 0-2625 00^ and 00672 H^O. C«6019 ; H«6-28. 
CigH^N^O^ requires C» 60-33 ; H = 6-16 per cent. 

Oxtdation in Presence of Sunlight, 

In the case of tartaric acid, it was shown {Brit, A$boo, Bepcri, 1895), 
that the oxidation to dihydroxymaleic acid may be brought about by 
atmospheric oxygen in presence of ferrous iron on exposure to 
sunlight; whilst in the absence of ferrous iron, or in the dark, 
practically no oxidation occurs. That the result was not due to 
atmospheric hydrogen dioxide or ozone was shown by using air 
previously purified by means of potassium iodide.* 

It is now found that a similar, although less rapid, result may be 
produced in the cases of glycol, glycerol, and erythritol. The oxidation 
products obtained with these substances in the manner described, 
all quickly restore the colour to a rosaniline salt previously bleached 

Digitized by VjOOQIC 


by sulphur dioxide, so that this test affords a delicate means of in- 
dicaUng the formation of such products. The experiments were made 
in the following way. 

A eolation of the alcohol was divided into three parts. To 1 and 2 
a litUe ferrous sulphate, in solution, was added and 3 was left blank ; 
1 was kept in a dark cupboard, and 2 and 3 were exposed to sunlight 
for some hours. The three samples were then tested with the rosani- 
line solution, and, in the case of each of the above-named alcohols, a 
strongly marked coloration was produced in 2, that is, in the one 
exposed to sunlight in presence of iron, but practically none was 
obtained in 1 or in 3. 

In the case of mannitol, dulcitol, dpc., the results are uncertain, since 
the rosaniline test is inapplicable, and the characteristic reactions of 
the products of oxidation, such as with phenylhydrazine, are scarcely 
delicate enough for the purpose. 

The greater part of the expense incurred in carrying out this and 
several previous investigations has been defrayed from grants kindly 
awarded by the Government Grant Committee of the Eoyal Society. 

II. — fi'Aldehydopropionic Add, CHO-OHa-CHg-COOH, 
arbd fi'Aldehydoisohutyric Add, OHO-CH2-CH(CH3)'COOH. 

By W. H. Pb&kih, jun., and C. H. G. Spbankling. 

Some time since (Trans., 1896, 60, 162), a paper was published by one 
of us in conjunction with Messrs. W. H. Bentley and E. Haworth 
which had for one of its objects the discovery of a method for intro- 
ducing tbe group 'GH^'CH^'OH into organic substances, a synthetical 
process which, if it could be easily carried out, would be very valuable 
as a means of forming ring compounds. It was found that the action 
of glycol ehlorhydrin, Cl'CHij'CHg'OH, on the sodium compounds of 
sabstanees like ethylic acetoacetate, ethylic malonate, and their 
derivatives, did not, except in isolated cases, yield the desired result ; 
ultimately, however, a method was devised which in the few cases tried 
gave fairly satisfactory results, and which may be briefly stated in the 
form of an example thus. 

/3-Bromethyl phenyl ether,G5Hg«0*GH2'GH3Br,wasprepared by acting 
on sodium phenoxide with ethylene bromide, and this, when digested 
wiihthe sodium derivative of ethylic methylmalonate, yielded ethylic 
y-pifaenoxyethylmethylmalonate , 

CeH,-G-CH,-CH,-0(OH3)(COOG2H5) 2. 
The add corresponding to this ethereal salt, when heated at 100^, loses 

„.gitized by Google 


one molecule of carbon dioxide with formation of y-phenozyethyl- 
methylacetic acid, C^H5-0-CHj-0H,-CH(CH8)-COOH, from which 
hydrobromic acid eliminates the phenyl group and forms y-bromethyl- 
methylacetic acid, CH2Br*OH3-CH(CH3)«OOOH. This brom-acid when 
boiled with sodium carbonate, gives the sodium salt of hydrozyethyl- 
methylacetic acid, OH-CH2«OH2-CH(CH3)*COONa, from which, on 

acidifying, methylbutyrolactone Y^ ^ Y **8^ ig at once 

obtained. When this method was tried in more complicated synthetical 
experiments, it did not work well, partly owing to the number of 
operations involved, but* principally on account of the smallness of the 
yield obtained in some of these operations. For these reasons, experi- 
ments were made with a view to discover a more direct method, and 
ultimately we found in hromaceUUf (C2H50)2CH'CH2Br, a substiince 
which seems likely to answer the purpose. 

Bromacetal reacts readily with the sodium derivative of ethylio 
malonate, yielding ethylic cteetalmaloncUe according to the equation 

(C,H50)30H-CH,-CH(COOC,Hj), + NaBr. 

This ethyiic salt, which distils without decomposition at 151 — 164^ 
(15 mm.), yields, on hydrolysis, the corresponding (iceUUmalanie add^ 
(C2H5O)2CH-CH2-0H(COOH)2,and this, when heated with water at 180*^ 
is decomposed into alcohol, carbon dioxide, and fi-aldehydoprapumie 
acid, {CjH50)2CH-CHj-CH(COOH)2 + H^O - CHO-OHj- CH,- COOH + 
2C2Hg'OH + COj. j3-Aldehydopropionic acid is a new and very inter- 
esting substance, since it is the " half-aldehyde " of succinic acid and 
belongs to the elate of aldehyde acids of which, iso far, very few have 
been prepared. Its properties show that it is a true aldehyde, and not 
a hydroxymethylene compound of the formula 
it therefore, does not belong to the class of substances which Olaisen 
has investigated with such brilliant results. Aldehydopropionic acid is 
an almost colourless liquid which reduces Fehling's solution and gives a 
violet coloration with a solution of rosaniline hydrochloride decolorised 
by sulphurous acid. It is slowly oxidised in contact with air, rapidly 
by nitric acid, with formation of succinic acid, and when reduced with 
sodium amalgam it yields buiyrolactane^ 

CHO-CHj-OHg-OOOH-OH-CHg-OHg-CHj-COOH- 9^'^^* ^. 

When boiled with caustic soda solution in a flat basin, aldehydopropionic 
acid undergoes a most interesting change, yielding small quantities of 
terephthcUie aeid, the dihydroterephthalic acid, which may be assumed 
to be the first product of the condensation, being oxidised to ter»- 

Digitized by VjOOQIC 


phthalie add hj the action of the air, 


oooh- chj- ch,- cho - ooohc • ch,- ch 

= oooh-c-ch:ch 

and this is, so far, the only experiment which has been instituted with 
the object of testing the valne of aldehydopropionic acid i^ conden 
sation experiments. 

The action oi bromaoetal on sodium compounds is probably a general 
onoi since we have found that the reaction proceeds equally well when 
the sodium derivatiYe of ethylic methylmalonate is substituted for that 
of ethylic malonate in the above experiments. 
The eihylie €teetalmethylmalanale, 

thus obtained yields aeeUdmethyhnalonie acid, 

on hydrolysifl, and this, when heated with water at 180^, is converted 
into /3-aldehydoisobutyricacid, CH0-CHg-CH(CF8)-CX)0H, a Uquid acid 
which, on oxidation, yields methyl succinic acid, 
Further experiments on the action of bromacetal on the sodium 
compounds of ethereal salts are in progress. It should be mentioned, 
in conclusion, that C. Harries (Ber., 1898, 31, 42) obtained a substance 
which is probably the methylal of the half-aldehyde of succinic acid, 
(CHgO)jCH'CH2-CHj-COOH, by the action of sodium hypobromite on 
levulin methylal, (OH30)2CH-CH3'GH,*CO-GH3, but he does not 
H^pear to have further investigated this substance. 

Aeiicn qf Bramaceial on the Sodium Derivative of Ethylic malonate. 
Formaiuyn of Ethylic AcetalmaU(mate,(0fifi)filI'Cn^'CR{COOC^U^\. 

Ethylic aoetalmalonate is conveniently prepared as follows. 
Sodium (14*2 grams) is dissolved in absolute alcohol (170 grams) and 
the scdution, while still warm, is mixed with ethylic malonate (100 
grams) and bromacetal (80 grams) and the mixture, inclosed in sealed 
tubes, is heated at 130^140° for 4 hours; when as much alcohol as 
possible has been removed from the product by distillation on the water 
bath, water is added to the residue and the precipitated oil extracted 
several times with ether. The ethereal solution is washed, well dried 
over calcium chloride, the ether distilled off, and the residual oil f rac- 
tionated under reduced pressure (16 mm.). More than half passes over 
below 145*^ and consists of a mixture of unchanged ethylic malonate 
tad bromaoetal, whilst the fraction distilling between 145° and 165° 

Digitized by 




contains the ethylio aoetalmalonate. The oil boiling below 146^ 
(15 mm.), and which was assumed to contain about 50 per cent, of 
bromacetal, was again heated in sealed tubes with the caleolated 
quantity of the sodium derivative of ethylio malonate, the temperature, 
however, being now allowed to rise to 160^ 

In this way, practically the whole of the bromaeetal was converted 
into crude ethylic acetalmalonate boiling at 145 — 166° (15 mm.), and 
from this fraction the almost pure ethereal salt could be obtained by 
repeated fractionation as a colourless oil of a peculiar and not unplea- 
sant odour, and boiling at 151—154'' (16 mm.), or at 166—168'' (at 
26 mm.). Analysis.* 

01821 gave 0-3714 CO2 and 01418 H,0. C=55-62 ; H-8-65. 
01640 „ 0-3360 CO3 „ 0-1282 H^O. 0=55-87 ; H- 867. 
(02H5O)j0H-0H,-CH(COOO2H5)2 requires = 56-50; H = 869 per cent. 

Acetalmalomc Acid, (C2H50)3CH-CH2-CH(C(X)H)3. 

Acetalmalonic acid is obtained by hydrolysing its ethereal salt 
with alcoholic potash, but the operation has to be carefully performed, 
since prolonged boiling with the alkali decomposes the acid with 
elimination of the group *CH2*CH(OC2H5), and formation of malonio 
acid. It is not clear how this decomposition takes place, but there can 
be no doubt as to the formation of malonic acid, since, in one instance, 
when a quantity of this acid was obtained melting at 131°, an analysis 
was carried out with the following results. 

0-2687 gave 03480 COg and 00972 H,0. 0= 35-32 ; H = 426. 
0Hj(C00H)2 requires = 34-63; H«3-86 per cent. 

It was very soluble in water, and when heated decomposed with 
evolution of carbon dioxide and formation of acetic acid. 

If, however, the action of the potash is only allowed to proceed for 
a short time, hydrolysis takes place normally, and a good yield of 
acetalmalonic add is obtained. After many experiments, we found 
that the following process gave the best results. Ten grams of the 
pure ethereal salt is mixed with a solution of 6 grams of pure potash 
in alcohol, and the mixture heated on the water-bath for 10 minutes, 
the alcohol is then rapidly driven off on the water -bath and the cold 
residue mixed with an excess of dilute sulphuric acid and extracted 
repeatedly with ether. The ethei-eal solution, after drying and 
evaporating, deposits a thick oil which shows no sign of orystalliBin^y 
even after standing for some days over sulphuric acid in a vacuum. On 

* The numbers obtained are slightly lov, on accoant of the oil containing traces 
of bromine, which it was found impossible to remove by fhustionation. 

Digitized by VjOOQIC 


ifiilysis^^it gave numbers agreeing approzimatelj with those required 
{oraoetahnalomc acid. 

L 0115O gave 0-2040 CX)2 and 00722 H^O. 0=4915 ; H- 697. 

n. 01779 „ 0-3177 OOjj „ 01200 H^O. = 48-72; H = 7-47. 
ffl. 0-1974 „ 0-3529 OOj „ 01227 H,0. = 48-75 ; H = 6-91. 
IV. 01537 „ 0-2719 002 „ 00915 H^O. = 4826 j H = 6-61. 
(000H)^CH.-CH,'OH(OO2H5)j requires = 4905 ; H«:7-27 per cent. 

Sth&rmU. — Aeetalmalonic add is very soluble in water, and if the 
aqueous solution is neutralised with ammonia, and silver nitrate added, 
a dense, white precipitate of the silver salt is precipitated, which, after 
washing first with water and then with alcohol and ether, gave the 
following results on analysis. 

0-2177 gave^ on ignition, 0109 2 Ag. Ag = 50*16. 
(C^50),OH'CH8' 0H(000Ag)3 requires Ag = 49-79 per cent. 

JDistilUuion qf AeeUUmalonie Aeid. — ^When this acid is heated in a 
fractionating flask, decomposition soon sets in with evolution of carbon 
dioxide, and the residue, after repeatedly fractionating under reduced 
pressure, yields a liquid which boils, apparently constantly, at 157 — 161^ 
(15 mm.), and which was at first thought to be acetalacetic acid 
(CjH40)2CH-OH2-OH3-OOOH. The analyses, however, show that this 
substance consists, for the most part, of aldehydopropionic acid 
(p. 16), elimination of alcohol having taken place during the distilla- 
tion owing, probably, to the unavoidable presence of small quantities 
of water. 

0-2658 gave 0-4740 00, and 01607 H^O. = 4863 ; H = 671. 
0-2514 „ 0-4372 OO2 „ 01437 HjQ. = 4853; H = 6-39. 
OOOH-OHj-OHg-OHO requires = 47 05 ; H-5'88 per cent. 
OOOH*CHj*OH2'OH(002H5)2 requires = 54-54 ; H = 909 per cent. 

In order to confirm this view, the liquid was heated with an equal 
quantity of phenylhydrazine for 10 minutes at 150° and poured into 
ether, when, on standing, a white, crystalline substance separated, 
which, after crystallisation from acetic acid, melted at 19P and gave 
the following results on analysis. 

0^)601 gave 01502 00, and 0-0350 HjO. » 681 7 ; H = 6-47. 
0-1258 gave 21-3 c.c. nitrogen at 19"^ and 768 mm. N = 19-42. 
CjeH^gN^O requires = 6809 ; H = 638 ; N = 1985 per cent. 

This substance^ on examination, was found to be identical with the 
condensation product formed by heating aldehydopropionic acid with 

* The four analyns given here were carried out with four di£ferent prepanitioDS 
of the add. 

Digitized by VjOOQIC 


phenjlhjdi'azine (see below), and thus it is probable that this aldehydo- 
acid was present in the oil obtained by the distillation of acetalmalonic 

fi'Aldehydopropumie acid, CHO-CHj-CHj-COOBL 

In order to prepare this substance, acetalmalonic acid is heated with 
about four times its weight of water at 180 — 190° for 4 hoursi the 
solution evaporated on the water-bath, and the residue allowed to stand 
over sulphuric acid in a well exhausted desiccator for 4 or 5 days. 

The yellow, oily residue thus obtained, on analysis, gave numbers 
agreeing approximately with thosev required for j^ldehydopropionio 

I. 0-2084 gave 0-3611 CX)g and 01127 HjO. = 47-40; H-6-01. 
U. 0-2124 „ 0-3675 CO, „ 01132 HjO. 0« 47-20 ; H= 5-92. 

III. 0-1703 ,,0-2928 00, „ 0-0890 H^O. 0« 46-89 ; H = 582. 

IV. 01379 „ 0-2300 OOj „ 00806 H^O. = 45-47; H = 647. 
V. 01576 „ 0-2649 00, „ 00909 H,0. = 45-90; H = 6-40. 

OHO'OHg'OHj'OOOH requires = 4705 ; H = 5-88 per cent. 

P-Aldshydopropionie cusid is a slightly brownish liquid which dissolves 
readily in water, freshly prepared, its solution reduces Fehling's solution, 
and produces a pink colour when mixed with a solution of rosaniline 
hydrochloride which has been decolorised with sulphur dioxide. When 
heated at 150° with phenylhydrazine for 10 minutes, condensation 
readily takes place, and if the product is poured into ether a white, 
crystalline substance separates on standing, which, after recrystallisa- 
tion from acetic acid, gave the following results on analysis. 

00590 gave 101 c.c. nitrogen at 18° and 758 mm. N*. 20'00. 

This substance, of which a fall analysis is given on p. 15, melts at 
192° and is evidently the phenylhydrazide of the phenylhydrazone of 
aldehydopropionic acid, O^Hg-NH^NIOH-OHj-OHj-OO-NH-NH-OeHg, 
which contains 19-85 per cent, of nitrogen. 

Oxidation of Aldehydopropionic Acid, FomiaUon qf Succinic Acid. 

When aldehydopropionic acid is left exposed to the air, it darkens in 
colour and gradually deposits crystals, ultimately being converted into 
a brown, pasty mass, which in contact with porous porcelain slowly 

* As this substance would not crystallise, analyses of each preparation were 
made, and some of these varied as much as 8 per cent from the theoretical ; we 
therefore wish it to be distinctly understood that we do not consider that the 
aldehydo-acid made in this way is pore. All the preparations contained a small 
amount of ash derived from the tube in which they were prepared ; this was allowed 
for in the analyses. 

Digitized by VjOOQIC 


beeomes a nearly oolourleBs, crystalline mass, the sticky, oily impurity 
being only very gradually absorbed. 

The crystals were purified by reorystallisation from hydrochloric 
add with the aid of animal charcoal, and in this way colourless plates 

were obtained which melted at 18P and had all the properties of 

Boecinic add. Analysis. 

01201 gave 01786 00^ and 00561 H^O. =i 4056 ; H = 518. 
00OH*Cn,*CHs-C0OH requires 0^40*68; H»508 per cent. 

This experiment shows that aldehydopropionio acid is slowly 
converted into succinic acid by the oxygen of the air, and this change 
takes i^aoe very rapidly when oxidising agents are employed. A small 
quantity of the aldehydo-acid, after being heated to boiling with dilute 
nxferieadd (20 per cent.) for some hours until no further oxidation took 
place, was evaporated repeatedly on the water-bath with the addition of 
small quantities of water, until a colourless, crystalline residue was 
left. This, after recrystallisation from hydrochloric add, melted at 
181 — 184° and consisted of succinic acid. 

0^598 gave 0-0898 OOj and 0-0279 HgO. C = 4096 ; H = 518. 
OOOH-GHj-CHs'COOH requires C » 40-68 ; H » 508 per cent. 

Adtcdion rf Aldekifdapropunne Acid. Formaiion of BuiyrolcustanSf 
CHj- OH,- CH, 

Flye grams of the pure aldehydo-acid were dissolved in 
water and treated with three times the calculated quantity of 4 per 
eent aodiom amalgam, carbon dioxide being passed through the liquid 
ukl the temperature kept below 10^ during the whole operation, in 
order to avoid, as far as possible, risk of polymerisation or condensa- 
tion ; after separating the mercury, the solution was made strongly 
Mad with sulphuric acid, heated to boiling for half an hour in a reflux 
i^iparstos, and then repeatedly extracted with ether. The ethereal 
solution, when dried and evaporated, deposited a colourless oil, which, 
after twice fractionating, boiled at 203—208°. On analysis, it gave 
niunbers agreeing with those of butyrolactone, which, according to 
Fittig and Boeder {Annalen, QSfJ. 1886, 22), boils at 206"^. 

1366 gave 0-2760 CO, and 00888 H,0. C = 55-51 ; H « 727. 
C^HjO, requires 0-55-81 ; H«6-97 per cent. 


Digitized by 



Action of Ccmatic Soda on P-Aldehydopropionie Add, Syntftens of 
' TorBphthalie Aeid. 

This synthesisi which is ezplained in the introduction to this paper, 
was carried out as follows. /3-Aldehydopropionic acid (10 grams) 
was dissolved in an excess of dilute sodium hydroxide and the solution 
evaporated in a flat glass basin nearly to dryness, water was then 
added and the solution again evaporated^ this operation being con- 
tinued during three days ; the concentrated liquid, after being acidified 
and allowed to stand over-night, deposited a small quantity of a 
brownish powder, which, on examination, was found to be crude tere- 
phthalic acid. This was purified by dissolving it in dilute sodium 
carbonate, boiling with animal charcoal until most of the colour had 
been removed, and then treating the solution at 0° with permanganate 
until the violet colour remained permanent for 2 minutes ; the filtrate 
from the manganese precipitate was coDcentrated, and, while still hot, 
acidified with hydrochloric acid ; the colourless, crystalline precipitate, 
which separated rapidly, had all the properties of terephthalic acid. 
It was almost insoluble in water and ether, did not melt at 270% and 
on heating in a small test tube it sublimed apparently without melting. 

The small quantity of acid remaining (about 0*2 gram) was converted 
into its methylic salt by Baeyer's {Annalen, 1888, 245, 140) method by 
heating with phosphorus pentachloride, mixing the product with 
methylic alcohol, and purifying the crystals which separated, by re- 
crystallisation from methylic alcohol, with the aid of animal charcoal ; 
the colourless plates thus obtained were very sparingly soluble in 
methylic alcohol and melted sharply at 140% the melting point of 
the methylic salt of terephthalic acid. A specimen of the methylic salt 
prepared from pure terephthalic acid was found to be identical with 
the synthetical substance in every respect, moreover, an intimate 
mixture of the two preparations melted sharply at 140^. These experi- 
ments prove conclusively that the acid formed by the action of sodium 
hydroxide on /3-aldehydopropionic acid is terephthalic acid, and it is un- 
' fortunate that, in spite of a number of experiments, we have been 
unable to devise a better method for the condensation of the aldehyde, 
as the yield of terephthalic acid obtained was certainly not more than 
5 per cent. 

P-AldekydoUobutfjric Acid, CHO*OH2*OH(OH8)-OOOH. 

The first step in synthesising this substance was to prepare ethylie 
acetalmethylmalonate, and this was readily accomplished by heating 
the sodium derivative of ethylie methylim^ionate with bromacetal 

Digitized by VjOOQIC 


ttuder the conditions already described in the case of ethylic acetal- 
malonate. The ethjlio malonate employed in these experiments 
was made by the etherification of pure methylmalonic add; the 
quantities of the sabstanoes nsed in the synthesis of ethylic acetal- 
methylmalonate were, 

Ifthylic methylmalonate 54 grams. 

Sodinm.... 7*2 ,^ 

Bromaoetal * 40 „ 

After isolating the product in the way described in the case of ethylic 
aoetalmalonatey the crude oil was submitted to careful fractionation 
and the fraction 166^(26 mm.), which consisted of nearly pure ethylic 
aoetalmethylmalonate, was analysed with the following result. 

0-1491 gave 0*3056 00, and 01204 H^O. O«67'20; H-8-97. 
{0AO)jCH-CHj-CH(OHsXCOOC^5)2 requires 0-57-92; H- 8-96 per cent. 

AcetalmMsfhMhnie Aeid, (0^oOVOH-GH2-OH(OHs)(OOOH),. 

This acid, which was prepared by the careful hydrolysis of its 
ethereal salt, is a colourless syrup readily soluble in water. 

01965 gave 0-3624 00, and 01404 H,0. « 50*30 ; H » 7*94. 
Oi^igO^ requires « 51 -28 ; H = 769 per cent. 

The tilver aaUf prepared by precipitating a neutral solution of the 
ammonium salt with silirer nitrate, is a white, amorphous powder, 
which readily darkens when exposed to light. 

0-2026 gave, on ignition, 0*0980 Ag. Ag » 48*37. 

O^oH^^AgjO^ requires Age 48*19 per cent. 

Aldehydoisobutyrie add, obtained by heating acetalmethylmalonic 
add witii water at 180^ for 4 hours, and evaporating the liquid on a 
water bath, is an oil which, after standing for some days over sul- 
phuric acid in a vacuum, was analysed with the following results.* 

0-1933 gave 03653 00, and 01238 H,0. «: 51-53 ; H ^ 7*28. 
CHO-OH,-OH(0H,)-0OOH requires 0-51*72 ; H-6-89 per cent. 

This Bubstanoe is very similar to aldehydopropionio add in its pro- 
perties, and its constitution is proved by the fact that, when oxidised 
with nitric acid, it yidds fnUh/^fhtteoifHe add ; this, after recrystallisa- 
tion, melted at 110 — 112°, and gave the correct numbers on analysis. 

0*2142 gave 0*3558 00, and 01288 H,0. = 45-30 ; H - 6-40. 
C00H-OH(OH3)-OH,'0OpH requires 0-45*46; H-606 per cent. 



«.. I ■ ■ '■ ■ ■ i « I I Ml ■ I ii » ii I ■ «t »m ■ , . I I . II ■ m 

* See footnote, ik 16. 



III. — CannahinoL Part I. 

By Thomas Barlow Wood, M.A. ; W. T. Nkwton Spivet, M.A. ; 
Thomab Hill EAiSTBRriSLD, M.A., Pb.D. 

In a paper oommunicatecl to the Society in 1896 (Trans., 1896, 69, 
539) the authors, under the name of '< cannabinol," described a 
physiologically active substance which they had isolated from 
** diaras/' the exuded resin of Indian hemp. From the constancy of 
composition of a number of preparations of this substance obtained 
from different samples of '* charas/' it was believed to be a definite 
chemical compound of the formula C^^B^fi^ i ^^ conclusion seemed 
to be justified by the determination of the molecular weight, and by 
the examination of several derivatives. Since then, the authors have 
further examined cannabinol, and have found that it is a mixture of 
at least two compounds having similar physical characters. One of 
these, of the formula C^^^O^ has been isolated, and it is proposed to 
retain the name cannabinol for this compound. 

During the progress of this investigation, a note on oxycannabin by 
Messrs. Dunstan and Henry appeared in the Proceedings of the Society 
(Proc., 1898, p. 44), in which the formula O10H10NO4 and CjgH^OAc 
were assigned to oxycannabin and acetylcannabinol respectively. 
FroBi the description of these substances, there can be no doubt that 
they are identical with those we have obtained, but the reeultfl 
of our analyses correspond with the formulas C^jH^iNO^ for oxy- 
cannabin and O^^^fi^'^i^fi ^^^ acetylcannabinol. These formula 
are confirmed by molecular weight determinations, and by analyses of 
many derivatives. Dunstan and Henry {loe. eit.) state that, on 
oxidising cannabinol with nitric acid, normal butyric acid is formed ; 
we can confirm this statement, with the addition that larger quanti- 
ties of normal valeric and caproic acids are produced at *the same time. 

As the present paper deals mainly with the substances produced 
by the breaking down of the cannabinol molecule, the authors have 
only described such of its reactions as suffice to show that it is a true 
ishemioal compound. An account of the reactions of cannabinol, 
together with a more complete examination of several of the com- 
pounds described below, will shortly be brought before the Society. 

When crude cannabinol (the red oil obtained by fractionating 
alcoholic charas extract under diminished pressure) is treated with 
nitric acid under certain conditions, it yields a yellow, crystalline 
substance, the analysis of which corresponds with the formula 
C^jHigN^jOs (Proc., 1898, p. 66), but determinations of the molecular 
weight show that a higher formula is required. As it has aoidio 

Digitized by VjOOQIC 


propratieBy Baits were inrepared and analysed, and their composition 
proTes that the formula is C^{H^fig^ which agrees with the molecular 
weight determinations. It is readily reduced by boiling with 
hydriodio aoid and phosphorus, yielding the hydriodide of a base ; 
the latter has not as yet been isolated, owing to the readiness with 
which it oxidises. As the formula Cg^H^gNgOg represents the trinitro- 
deriyatiye of a compound, G^^^fiv the presence of the latter in crude 
cannabinol is (urobable, and this is definitely proved by the isolation 
of the acetyl derivative, C^B.^O^^CJS^Oy from a crude cannabinol. 

After the separation of the 'crystalline acetyl derivative from the 
products obtained by the acetylation of crude cannabinol, an oily 
residue was left amounting to more than three-quarters of the acetyla- 
tion product, and from it by treatment with acetic anhydride, no 
other crystaUine acetyl derivative could be obtained ; it appears to 
oontain Uie acetyl derivatives of one or more substances with a lower 
percentage of carbon than GgiHggOj. 

" Crude cannabinol " is, therefore, a mixture of cannabinol, CjiH^Ogy 
with one or more compounds probably of lower molecular weight. 

The fact that cannabinol forms an acetyl derivative proves that it 
contains a hydrozylic group, and the failure of all attempts to 
obtain an ethereal salt from the trinitro-compound referred to above, 
makes it probable that the acidity of the latter is not due to the 
presence of a carbozyl group^ but to the influence of the nitro-groupe on 
the hydroxyl group. 

On oxidising the trinitro-derivative by boiling it for several hours 
with fuming nitric acid, and pouring the product into water, a yellow, 
flooculent precipitate was deposited ; this was filtered ofC, and the 
add filtrate steam distilled ; the distillate was found to contain normal 
butyric, valeric, and caproic adds. The normal butyric acid was 
identified by its ealdum salt, the normal valeric and caproic adds by 
their anilides. By direct oxidation of crude cannabinol with 
nitric add, the same fatty adds were obtained, and caproic acid 
was aieo obtained when potassium permanganate or chromic acid 
mixture was employed. 

The yellow, flocculent precipitate, on crystallisation, gave a mixture 
of oily adds containing nitrogen, and a pale yellow, crystalline com- 
pound, the oxycannabin of Bolas and Frauds (Chem, News^ 1871, 24, 
77) ; to this they gave the formula O^fl^^ji^jt l>ut the results of 
their analyses agree equally well with the authors' formula G^jHi^NO^. 
Dunstan and Henry, in their note (Froc., loc. cit.), state that oxy- 
cannabin " does not dissolve in aqueous alkalis unless warmed with 
them in a closed tube. By addifying the resulting solution, an acid 
is predpttated which is at present under investigation. Oxycannabin 
would, therefore, appear to be a lactone." The present authors find, 

Digitized by VjOOQIC 


however^ that ozyoannabin disBolves in aqueous oanstic soda if boiled 
iHth it for a few minutes, and that unaltered ozyoannabin is preeipi- 
tilted on addifioation. They have proved it to be a lactone by the 
preparation and analysis of salts of the corresponding ozy-aoid, but 
this ozy«acid has not been isolated, since it is at once re-converted 
into the lactone when set free from its salts. In fact, its tendency 
to undergo this change is so great that, on treating the silver salt 
With ethylio iodide, the lactone is obtained instead of the ethylie 

On oxidation with dilute nitric add at 186°, ozyoannabin yields a 
sparingly soluble nitro-lactonic acid of the formula GjQHgNO^'OOOH 
together with a very soluble tribasic acid, C^H^NOg. 

On reduction, ozyoannabin yields a compound, Oi^H^gNO^t which 
corresponds to the reduction of a nitro-group in ozyoannabin to an 
amido-group. On this evidence, the authors propose for ozycannabin 
the name nitrocannabino-lactone, and for the reduction product the 
name amidocannabino-lactone. 

The amido-kctone is readily diazotised, and an attempt was made 
t6 prepare cannabino-lactone directly from it by Friedlander's method, 
but it was found more convenient to prepare iodocannabino-lactone, 
OjjHjjIO], by adding potassium iodide to the diazotised solution, and to 
riiduce this with sodium amalgam in alkaline solution. The oily canna- 
Mno-lactone, Oj^HijOg, thus obtained, was converted into colourless 
crystalline, cannabino-lactonic acid, Gj^H^oO^, by boiling it with an 
alkaline solution of potassium permanganate ; this action corresponds 
to the oxidation of a methyl group to a carbozyl group. On reduction 
With hydriodic add and phosphorus, the lactonic acid yields a dibasic 
Add of the formula O^jH^^O^ thus affording confirmation of its lactonio 

From the fact that a methyl group, both in cannabino-lactone 
ted in its nitro-derivative, is ozidised to a carbozyl group, and from 
(he behaviour of nitrocannabino-lactone on reduction, and subsequent 
diazotisation, the presence of a benzene nucleus in cannabino-lactone is 
probable. In order to confirm this, and, further, to ascertain the 
Mructure of the lactone ring, the lactonic acid was fused with caustio 
tK>ta8h ; in this way, isophthalic add was obtained, and identified by 
conversion into its methylic salt. However, when cannabino-lactone 
Was fused with potash, metatoluic acid was formed, together with iso- 
phthalic acid. The formation of these two adds definitdy proves 
the presence in cannabino-lactone of a benzene nucleus, with two 
Aide chains in the meta-position relatively to each other, one of 
these chains bdng a methyl group. The second side chain must 
evidently contain the lactone ring, and on the assumption that it is a 
y-lactone, which from its great stability is probable, there appear to 

^.gitized by VjOOQ__ % 


be mily three possible formula, namelj, those of the thiee metatolyl- 

CH,-9,H, CH,-9,H, CH,-9,H, 


Synthesee of these three lactones are at present in progress, with a 
view to deciding which of them is identical with cannabino-Iactone. 

The following table shows the relationships of the -oompoands 
described in this paper. 

Gannabinol, O^^H^O,. 
Aoetylcamiabinol, C^HgjOj-CXJCH,. 
Trinitrocannabinol, 0„Hj3fe^O,)8' Og. 

Oannabino-lactone, (l)CH3-CgH^<^^[»>CO(3). 

Nitrocannabino-lactone(oxjrcannabin), (1) ^0>C^U^<S^^i^><X) 

Amidocannabino-lactone, (1) ^«>CeH3<^^i>C0 (3). 

lodocannabino-lactone, (1) ^^p>C^M^<S!^^i^>00 (8). 

Oannabino-hictonio acid, (1) COOH- C3H^<^i>00 (3). 

Nitrocannabino-lactonic acid, (1) ^Q>CJ^^<S^i>00 (3). 
Carbozyphenylbatjric add, OOOH-OgH^^CjH^-OOOH (3). 


Triniiroeannabinols ^/^^^{ISO^fiy — This compound is produced on 
adding fuming nitric acid (5 cc), drop by drop, to a well cooled solu- 
tion of crude cannabinol (8 grams), dissolved in glacial acetic acid 
(18 cc), the temperature being kept down by immersing the flask in 
ioeeold water. After standing for several days, the crystals are col- 
leeted; the yield is 20 per cent. (1*6 grams) of the crude cannabinol 
used. The compound is easily soluble in benzene, phenol, alcohol^ and 
ether ; it is also readily soluble in hot, but only sparingly in cold, glacial 
acetic acid, which is the most convenient solvent for its recrystallisa- 
tion. It crystallises in bright yellow plates, which, when quickly heated, 
melt at 160^ (uncorr.) with some decomposition. 

Four preparations were analysed 

A and B, purified through the ammonium salt. C. Five times re- 
crystallised from glacial acetic acid, D. Sample 0, once recrystallised 
from alcohol. 

Digitized by VjOOQIC 


^ f 0-1417 gave 02935 C», and 00678 H,0. 

' 1 0*1235 „ 10 CO. moist nitrogen at 18^ and 760 mm. 
B. 0- 


B. 01676 „ 0-3480 00, and 0-0835 HjjO. 

•1860 „ 0-3855 00, „ 00913 H,0, 

1-1440 f, 11-5 C.C. moist nitrogen at 21^ and 756 mm. 

D. 0-1527 „ 0-3180 00, and 0-0750 H,0. 

Calculated for Foand. 

CnH,|N,Oa. A. B. ' a D. 

0-56-6 percent 56*5 566 565 56-8 

H« 5-2 „ 5-3 5-5 54 55 

N= 9-4 „ 9-3 — 90 — 

The molecnlar weight was determined by the freezing point method 
in benzene and in phenol. 

In benzene (i) 0*5220 gram, dissolved in 20 grams of benzene^ lowered 
the freezing point by 0-312"^. 

(ii) 0*9513 gram, dissolved in 20 grams benzene, lowered the freezing 
point by 0-570°. 

In phejiol (iii) 0*4733 gram, dissolved in 20 grams phenol, lowered 
the freezing point by 0-410°. 

(iv) 1-0023 grams, dissolved in[20 grams phenol, lowered the freezing 
point by 0-900° 

Calcalated for Found. 

CiH„N,Os. (i). (u). (iii). (iv). 

Molecular weight -445 409 409 427 412 

Satae. — ^The compound has acidic properties, forming salts of potas- 
sium, sodium, ammonium, and silver, all of which are bright yellow, 
crystalline compounds sparingly soluble in water, easily in alcohol. They 
are best crystallised by diluting their hot alcoholic solutions with water. 
The potassium and sodium salts are explosive, the silver salt is not 

Sodium Salt. — The most soluble of all the salts examined, the 
saturated solution at 15° containing 1 part of salt in 120 parts of 
solution. It is prepared by dissolving the acid in excess of alcoholic 
soda solution and diluting with water. Two samples were analysed* 

0*2130, at 160°, lost 00290 H^O and gave 0-0265 Na,S04. Na» 3*9 ; 

0-1545 gave 00205 Na,S04. Na=4-3. 

NaO^jH^^NsOg + 4H2O requires Na - 4*3 ; H^ - 13*4 per cent 

Fotaeeium salt, prepared in the same way as the sodium salt, 
equires for solution 2000 parts of water at 15°, and 500 at 100°. 

. 0*1210 lost no weight at 150°, and gave 0*0220 K^^. E»8-2. 
0-1300 gave 0-230 E^SO^; K-7-9. 

KOjiH^^gOg requires K»8*l per cent. 

Digitized by VjOOQIC 


Ammonium tali may be prepared by dissolving the aoid in boiling 
ammonia solution, when the salt crystfdlisee out on cooling, or by dis- 
solving the acid in excess of alcoholic ammonia and diluting with 
water. It requires for solution 1600 parts of water at 15% and 200 
at 100°. 

01170 gave 0-2330 OOj and 00605 H^O. = 543 ; H = 6-7. 
0-1312 ,, 13*3 C.C. moist nitrogen at IS'' and 758 mm. N» 11*7. 
0-4190 „ on distillation with sodajNHg = 9-7c.c.N/10 HCL NHg = 3*9. 
01320 „ 0-2640 00, and 00690 H^O. = 54-5 ; H = 5-8. 
0-1695 „ 170 C.C. moist nitrogen at 19° and 750 mm. N = 121. 
NH4-OaH2jN808require80«54-5;H=:5-6;N = 121 ;NH, = 3-7 per cent. 

Sther 9aU, prepared either by adding silver nitrate solution to the 
Bolntion of the sodium salt, or by boiling the alcoholic solution of the 
add with excess of silver carbonate. A. Was prepared from the sodium 
salt and recrystallised from alcohoL B. From the acid and silver 
carbonate. is B recrystallised. 

(0-1900 gave 0-3180 OOg, 00715 Hfi, and 00362 Ag. 0»45-6; 
A.-J H = 4-2;Ag=191. 

V 0*1829 gave 11-2 c.c. moist nitrogen at 16° and 766 mm. N » 7*2. 

B. 01580 „ 0-2673 OOj, 0-0586 HjO and 00295 Ag. = 461; 

H=41;Ag = 18-8. 
(0-1850 gave 0-3090 CO2, 0-0675 Kfi „ 00357 Ag. 0-45-5; 

C. < H = 41;Ag = 19-3. 

' 0-2600 gave 16*2 c.c. moist nitrogen at 18° and 758 mm. N = 7'2. 
AgCjjHjjNjOg requires = 45-6 ; H = 4*0; N = 76; Ag = 194 per cent. 

BaducUon i^f TrifUH'ocaMnalnnol, — ^Trinitrocannabinol is reduced by 
bailing with hydriodic aeid and phosphorus in acetic acid solution ; 
3;7 grams of trinitrooannabinol were dissolved in 50 c.c. of glacial 
acetic add^ and boiled for Ij^ hours with 25 c.c. of hydriodic acid 
spu gr. 1*69 and 4 grams of yellow phosphorus; the almost colourless 
solution was then filtered from the excess of phosphorus, and distilled 
down to one-third its volume. The pale yellow crystals of hydriodide 
which were deposited on cooling were collected, washed with acetic 
add, and dried in a vacuum over solid potash. 

Attempts to isolate the base failed, as when set free from the 
hydriodide it is at once oxidised, with fprmation of coloured products. 

The hydriodide was analysed, and gave the following numbers, 
0=41-2; H=5-2; 1=413; N = 50 per cent. 

The authors do not feel justified in making any statement as to the 
constitution of the product of reduction until they have isolated the 
base and made a further examination. 

Aceiylcannabinol^ O^H^gO^-OO-OHg.— Orude cannabinol is readily 

Digitized by VjOOQIC 


aoetylated by boiling it with aoetic anhydride or with acetic chloride. 
After distilling off the excess of acetic anhydride and rectifying the 
residue under diminished pressure, an oil, lighter in colour and more 
mobile than the original crude cannabinol, is obtained; this some- 
times, on stfinding, but, better, by dissolving it in alcohol and cooling 
the solution to 0^, deposits a white, crystalline substance, which can be 
purified by crystallisation from alcohol, light petroleum, or acetic acid ; 
it melts at 75^, and is evidently identical with the * aoetylcannabinol, 
OjgH^gOAc,' mentioned by Dunstan and Henry {loc. eU,y 

Analysis indicated as the simplest formula OjgHjgO^ (Proc., 1898, 
66), but determination of the molecular weight and the results of 
saponification show that the higher formula, C^sH^gOg, must be adopted 
(IVoc., 1898, 163). 

The following samples were analysed. A. Recrystallised from 
alcohol. B. Four times recrystallised from alcohol. 0. Becrystal- 
lised from light petroleum and then from acetic acid. 

•1637 gave 04420 00^ and 01101 H^O. C = 78-4 ; H= 79. 
•1118 „ 0-3208 OOj „ 00820 HjO. C-78-3 ; H = 8-l. 
^ f 01250 „ 0-3695 COj „ 00914 HjO. C = 78-4; H = 81. 
10-1290 „ 0-3692 CO2 „ 00928 HjO. 0-781 ; H- 8-0. 
0. 01860 „ 0-6315 COg „ 0-1330 HgO. 0-77-9 ; H- 8-0. 
^16^18^2 requires = 78-3 ; H = 7'8 per cent. 
^21^25^2* OO-OHj requires = 78-4 ; H2O = 80 per cent. 
O^gH^gOAc (Dunstan and Henry) requires = 80-6 ; H=8*7 per cent. 

The molecular weight was determined by the freezing point method, 
in glacial acetic acid and in benzene. 

In gkbciai ctcetie cteid, — (i) 0*2043 gram, dissolved in 20 grams of 
glacial acetic acid, caused a depression of 0*123°. 

(ii) 0-6052 gram, dissolved in 20 grams of glacial acetic acid, caused 
a depression of 0*310°. 

In 60fi«0ne.*— (ili) 0*1877 gram, dissolved in 17*04 grams of benzene, 
caused a depression of 0*140°. 

(iv) 0-3032 gram, dissolved in 17*04 grams of benzene, caused a 
depression of 0*240°. 

1 0-1 

Calculated for 

Calcnlated for 




(i). (ii). (iii). (iv). 

Mol. wt. = 230 


324 318 386 365. 

The percentage of acetyl was determined by saponifying the com* 
pound, either by boiHng with alcoholic potash, or by heating with 
it in a sealed tube at 130°. Water was added to the product, 

* For the molecular weight determination in benzene solution, we are in^^bted ta 
the kindness of Mr. H. Jfickson, B. A,, pownin^ College. 

Digitized by VjOOQIC 


and, after the alcohol had been boiled off, the solution was made 
strongly add with phosphoric acid, and steam distilled until the dis- 
tillate no longer had an acid reaction. The distillate was then titrated 
with aenu-normal caustic soda solution. 

1'6 gave acetic acid equivalent to 8'85 c.c. N/2 NaOH solution. 
^•0 n „ „ 52-4 c.c. 

Calcukted for Galcnlated for 

Ci^HuOi. CnHaOj'CO'CH,. Found.^ 

Afistyi-lS-T per cent. 12-2 11*9, 12*5. 

Cannabinoly OjiH^Oj. 

The residaes left in the distilling flasks in the above acetyl determi- 
nations were extracted with ether. The ethereal solution was dried 
over cakdam chloride, the ether distilled off, and the residue distilled 
onder diminished pressure, when practically the whole passed over at 
^^nnderSO mm. preesure; the distillate was an almost colourless oil, 
which, on cooling, set to a transparent resin. A second preparation 
was made, and both were analysed. 

01880 gave 0-6687 COg and 01446 HjO. 0«81-5 ; H = 8-6. 
0-1716 „ 0-5100 CX)2 „ 01300 HgO. C-8M;H=:8-4. 
CqHs^O, requires C»81-3 ; H = 8*4 per cent. 

The molecular weight was found by the freezing point method in 
glacial acetic acid solution. 

0-5876 gram, in 20 grams of glacial acetic acid, gave a depression 
of 0-380°. 

0-6260 gram, in 20 grams of glacial acetic add, gave a depression 
of 0-390°. 

Calculated for 



1. wt.-310. 

310, 313, 

The compound is optically inactive. 

Oitidaiion qf TfinitroccmncUnnol with Nitric Acid, 

One hnndred and twenty grams of trinitrocannabinol was dissolved 
m 300 0.C of hot, fuming nitric acid, and gently boiled for 5 hours 
^ a lefloz apparatus, more acid being added from time to time, until 
>A ^ 700 cc. had been used. On pouring the product into water, 
70 grams of a yellow, waxy precipitate separated, and from this 
36 grams of nitrocannabino-lactone (ozycannabin) were obtained on 
treatment with alcohol, Xbe filtrate from the above yellow precipitate 

„.gitized by Google 


smelt strongly of yalerio acid, and contained normal caproio, valerio, 
and butyric acids, possibly also propionic acid, together with non- 
volatile acids. Substantially the same products were obtained by 
the oxidation of crude cannabinol with nitric acid. 

Examination qf the Volatile Fatty Acide. 

The volatile acids were removed from the above-mentioned acid 
mother liquor by steam distillation, and the distillate neutralised 
with sodium carbonate and evaporated to dryness ; on acidifying 
with dilute sulphuric acid and extracting with ether, 12 grams of the 
mixed acids were obtained. These were fractionated with the follow- 
ing results. 

105— 160°= 2-2 grams. 170— 180°= 1-2 grams. 

160— 170°=0-9 „ 180— 195°=3-9 „ 

Above 195°= 1-86 grams. 

From the fraction 150 — 170°, normal butyric add was isolated by 
means of its calcium salt, which was purified and analysed, with the 
following results. 

00730 lost 0-0037 H^O at 135° and gave 0-0418 OaSO^. H3O « 7-8 ; 
(C4H90j)jCa + HgO requires HjO = 77. Ca = 172 per cent. 

The valeric add in the fraction 180 — 195° was separated as the 
valeranilide which melted at 60° (uncorr.), and on analysis gave the 
following results. 

00660 gave 01795 00^ and 00515 H^O. 0=74-2 ; H = 8-7. 
01998 „ 13*6 C.C. moist nitrogen at 20*5° and 767 mm. N = 7*8. 
CiiHijNO requires C = 74*6 ; H = 85 ; N = 79 per cent. 

As the anilide of normal valeric acid has not been described, it was 
prepared from a sample of normal valeric acid boiling at 186 — 186°, 
and was found to melt at 61° (uncorr.). Admixture of the anilide 
obtained from the fraction 180 — 195° with the anilide from normal 
valeric acid did not depress the melting point of the latter. The acid 
in this fraction is, therefore, normal valeric acid. 

The fraction boiling above 195° was converted into the anilide which, 
after many crystallisations from light petroleum, melted at 93*^ 
(uncorr.). The analytical numbers were slightly low for a oapro- 
anilide, and lack of material prevented further purification. KornoLal 
caproanilide melts at 95°, and since it has been shown that the lower 
acids belong to the normal series, it can hardly be doubted that the 
acid obtained from this fraction is normal caproic acid. 

In a similar examination of a much larger quantity of the fatty 

„.gitized by Google 


adds obtained by oxidation of crude cannabinol with nitric acid, the 
same acids were obtained. Normal butyric acid was identified by its 
caksinm salt, normal valeric acid by its anilide, and caproic acid by its 
silver salt. 

The fraction 150 — 170^ was also examined for isobutyric acid 
(b. p. 155'') by the method used by Y. Meyer and Hutzler (Ber., 1897, 
2519), but as no trace of acetonio acid could be detected after oxida- 
dation with potassium permanganate, isobutyric acid is evidently 

Gaproic add was also found in the product of the oxidation of 
crude cannabinol by chromic acid mixture, and by potassium per- 
manganate solution. 

In each case, the silver salt was recrystallised until the analytical 
results were constant. A, From chromic acid mixture oxidation. 
B. From potassium permanganate oxidation. C. From nitric acid 

f 01459 gave 0-1745 00^ and 00648 H^O. C = 325 ; H « 4-9. 
-^•101650 „ 00800 Ag. Ag = 48-5. 
B. 01070 „ 00515 Ag. Ag=i481. 
0. 0-2383 „ 01158 Ag. Ag = 48-6. 
CjHijCOjAg requires - 323 ; H = 4-9 ; Ag = 48 3 per cent. 

IfUrocai'mabmo4{icUma (oxycannabin), 

C„H„NO,-NO,-CeH,(OH,)<:^|^0 (3). 

(1) ^ 

The preparation of this substance from trinitrocannabinol has already 
been described ; it can be obtained more conveniently by the following 
method. Crude cannabinol is dissolved in three times its weight of 
glaeial acetic acid, warmed to 100°, and nitric acid (sp. gr. 1*42) slowly 
dropped in from a burette in the proportion of 1 c.a of nitric acid for 
each gram of cannabinol ; the solution is then boiled gently for half 
an houTy more nitric acid is added, and the boiling continued for 
8 — 10 hours, nitric acid being added whenever the oxidation slackens, 
50 grams of crude cannabinol require in all 300 — 400 c.c. of nitric 
aoid. The product is then poured into water, and the nitrocannabino- 
lacione separated as before 3 the yield is 15 — 20 per cent, of the crude 
eaonabinol used. When purified by repeated crystallisation from al- 
cohol, it is obtained in very faintly yellowish needles melting at 178° 
(unnnrr.), and is not changed by sublimation. On exposure to light, it 
gradoally assumes a reddish tinge. It is soluble in alcohol, acetic 
acid, benxefue, and concentrated nitric acid. These properties show 
that it is identical with the oxycannabin of Bolas and Francis 
(loe. et^.), and of Donstan and Henry {loe. cU.). Five samples were 

Digitized by VjOOQIC 

1 01 


1 0-1 


analysed. A. Purified by sublimation. B. Precipitated from the 
potassium salt. C. Crystallised from dilute acetic acid. D. From 
alcohol. £. From trinitrocaxmabinol, and orystailiaed from aloohoL 

A. 01478 gave 0-3225 CO, and 0-0663 Ufi. C = 59-5 ; H^SO. 

B. 01313 „ 0-2852 COg „ 0-0580 H,0. C = 69-2; H-4-9. 

C. 0-1195 „ 0-2615 COg „ 0-0582 H,0. C = 69-7; H^5-4. 
[01330 „ 0-2900 GO, „ 00605 H,0. C-=59-5; H = 61. 

'1615 y, 8-8 c.c. moist nitrogen at 2P and 772 mm. N «» 6*3. 

•1415 „ 8-0 „ „ 17° „ 764 „ N = 6-6. 

E. 0-1155 „ 6-5 „ „ 19° „ 750 „ N«=6-4.^ 

CjiHiiNO^ requires C = 597 ; H «= 5-0 ; N = 6-3 per cent 

CjjjjH^o^jOy (Bolas and Francis) „ C = 600 ; H = 5-0 ; N = 7 -0 „ 
CioHi^^04(Dunstan and Henry) „ C = 57-7;H = 4-8;N = 6-7 „ 

The molecular weight was determined by the f reeaing point method 
in gladal acetic acid. 

0*4313 gram, dissolved in 20 grams glacial acetic acid, lowered the 
freezing point 0-375°. 

Mol. wt. calc. for CuH^NO^ - 221. Found « 226. 

Nitrocannabino-lactone is insoluble in cold aqueous alkalis, but 
dissolves on boiling for a few minuteSi and is not precipitated on 
dilution with water; the addition of mineral acids, however, pre- 
cipitates it unchanged. Its salts are prepared as follows. 

FoUunum salt — This salt separates in slender, pale yellow needles 
on mixing saturated solutions, in absolute alcohol, of the lactone and 
caustic potash. 

• 01197 gave 0-0375 K^SO^. K = 14*l. 

CiiHijNOgK requires K- 14*0 per cent. 

Silver aaU. — ^Prepared by adding silver nitrate to an aqueous 
solution of the potassium salt. 

0-1322 gave 0-0409 Ag. Ag - 30-9. 

CnHjjjNOgAg requires Age«31*3 per cent. 

' An attempt was made to obtain the ethylic salt by boiling the 
silver salt with ethylic iodide ; silver iodide separated, but on extracting 
the product with boiling alcohol, nothing but unaltered nitrocaniiabixio- 
lactone was obtained. 

The lactone was treated with aqueous ammonia, sp. gr. 0-88, and 
with saturated alcoholic ammonia, both at the ordinary temperature, 
and in sealed tubes at 100°. In each case, nothing but the unaltered 
lactone could be recovered. Failing to obtain an amide, an at- 
tempt was made to prepare the anilide, but this also was unsuooeesf ul. 

Digitized by VjOOQIC 


Oxidaiion qf Nitro&mnabino-lactane wUh dilute NUric ^cu£.— Nitro- 
cannabino-lactone (4 grams) was heated with 25 per cent, oitric aoid 
(80 e.c) at 185° for 8 hours, and the liquid, on stirring, deposited a 
oolonrless, crystalline, sparingly soluble acid ; this, after recrystallisa- 
tion from hot water, melted at 229 — 230° (uncorr.). The weight of 
the recrystallised acid was I'l grams. 

Analysis shows that the substance has the formula Oi^H^O^, 
which may be deriyed from nitrocannabino-lactone by the oxidation of 
a methyl group to a carbozyl group, thus 

OOOH- CeHj(NOj)<^«j!>CO. 

^' { 0. 

This conclusion is strengthened by the fact that nitrocannabino- 
lactone, on oxidation with potassium permanganate in the cold, yields 
the same product. A. Was prepared by oxidation with nitric acid. 
B. With potassium permanganate. 

01053 gave 00360 H^O and 2033 CX),. C = 527 ; H = 38. 
'1328 „ 6 '5 cc. moist nitrogen at 21° and 750 mm. N = 5*5. 
/ 01725 gave 00612 H,0 and 03300 00,. = 522 ; H = 3-9. 
^' I 0-1700 „ 8-0 C.C. moist nitrogen at 20° and 765 mm. N = 5*4. 
CiiHjNO^ requires * 62-6 ; H = 36 ; N = 55 per cent. 

The acid, neutralised with ammonia, was converted into silver salt. 

0-1568 gave 00472 Ag. Ag ^ 30*1. 

Oi^HgNO^Ag requires Aga30'2 per cent. 

This acid must be regarded as the nitro-derivative of the cannabino- 
lactonic add described below. 

The nitric acid mother liquor, from which the sparingly soluble acid 
had separated, was evaporated to dryness, the residue dissolved in a 
very small quantity of cold water, filtered, again evaporated to 
dryness and extracted with ether. The syrup left on evaporating 
the ether gradually deposited crystals which were excessively soluble 
in water, alcohol, ether, glacial acetic acid, and ethylic acetate, but 
practically insoluble in bensene, chloroform, and light petroleum. The 
compound is most satisfactorily purified by recrystallisation from 
sUong hydrochloric acid, when it is obtained as a colourless, crystal- 
line powder, melting at 228 — 230° with much effervescence, but only 
a slight discoloration. 

Three preparations were analysed^ 

L 0-0948 gave 0-0188 HjO and 01478 COj. = 425 ; H =. 2-2. 

0-0760 „ 3-5 C.C nitrogen at 20° and 760 mm. N » 5*5. 
n. 01646 „ 0-0302 H^O and 0-2552 OOj. = 42-3 j H = 20. 
IDL 01775 „ 0-0325 HgO „ 02754 00^ = 42-3; H = 20. 
The fomuU C^HjNOjT^quires O = 42-3 ; H = 20 ; N = 5-5. 

Digitized by VjOOQIC 


The acid is tribasic^ for on titration with N/10 soda and phenol- 
phthalein, 0*065 gram required 7*5 c.c. for neutralisation. A tribasic 
acid of the above formula requires 7*65 c.c. 

A solution of the calcium salt of the acid was prepared by neutralisa- 
tion with chalky and from this, nitrate of silver precipitated the silver 
salt as a primrose-yellow powder, slightly soluble in hot water. The 
salt is feebly explosive. 

0-113 gave 0004 H^O and 0079 OO2. = 19-0; H = 0-4. 
01565 „ 0-1172 AgCl. Ag = 56-4. 

CgHgNOgAgg requires 0=18-7; H«0-3; Agc=56-2. 

No attempt has been made to elucidate the constitution of the acid ; 
it is, however, difficult to account for the formation of a tribasic acid 
of the formula O^H^NOg from nitrotolylbutyrolactone unless the acid 
is a hydrozyglyoxylic acid of the constitution 


Amidooannabino-lacUme, OuHuNHg- O^. — ^This compound can be ob- 
tained by the reduction of the nitrolactone by tin and hydrochlorio 
acid, or by hydriodic acid and phosphorus^ but the latter is more 

Ten grams of the nitrolactone dissolved in 40 c.c. of glacial acetic acid 
was boiled with 30 c.c. of hydriodic acid of sp. gr. 1*6 and 5 grams of 
phosphorus for 2 or 3 hours, the colourless solution on cooling deposit- 
ing crystals of the hydriodide of the base, and a further quantity of 
the base was obtained by pouring the strongly acid mother liquor 
into water. On dissolving the hydriodide thus obtained in boiling 
water, dissociation occurs, and the free base crystallises out on 

Amidocannabino-lactone crystallises readily from hot water in long, 
white needles melting at 119^ (uncorr.). Two preparations were 
analysed. A. Prepared by hydriodic acid and phosphorus. B. By 
tin and hydrochloric acid. 

•1260 gave 0-3200 00^ and 0-0770 H^O. = 693 ; H = 6-8. 
•1140 „ 0-2895 OOj „ 00720 H^O. = 69-2; H = 7-0. 
•1826 „ 11-5 C.C. moist nitrogen at 22"" and 770 mm. N = 7'2. 
B. 00795 „ 0-2005 OOj and 00500 HjO. = 68-8 ; H « 70. 
OuHi3^0j requires = 691 ; H = 68 ; N « 7-3 per cent. 

The hydriodide was also analysed. 

0-1957 gave 01420 Agl. I = 39^2. 

OjiHii08'NH3,HI requires 1 = 398 per cent. 

The platinoohloride was prepared by dissolving the base in strong 
hydrochloric acid and adding platinic chloride solution. Two specimens 
were analysed. 

Digitized by VjOOQIC 




01200 gave 0-0296 Pt. Pt = 24-7. 
0-2594 „ 00636 Pfc. Pt = 24-5. 

(CiiHiiOg'NH,)yH,Pt01e requires Pt = 24-6 per cent. 

The base was readily diazotised, and an attempt was made to 
prepare cannabino-lactone from it by Friedlander's method, bat no 
8atisfsct<»7 prodact could be obtained. 

Oaumbino-lactone was, howeyer, easily obtained through the iodo- 
laetone, whose preparation is given below. 

Iodoainnabino4aeUmB, CHj* CJBi^l<^^J^<K).—Ten grams of amido- 

cannafalno-lactone was dissolved in 25 c.c. strong hydrochloric acid, 
75 cc water added^ and the solution diazotised by the addition of 
4*5 grams of sodium nitrite dissolved in 15 c.c. of water, the tempera- 
ture being kept within a few degrees of the freezing point. 12*5 
grams of potassium iodide in 25 cc. water was then added, and the 
mixture heated on the water bath until evolution of nitrogen ceased ; 
the acid liquor was then poured off, and the solid residue, after treat- 
ment with solution of sodium thiosulphate to remove free iodine, was 
repeatedly crystallised from dilute acetic acid. The yield was 10 grams. 
The iodolactone forms almost colourless crystals melting at 137*5° 
(uncorr.), insoluble in water, but easily soluble in alcohol and acetic 
scid ; it sublimes very readily. 

0*2370 gave 0*3825 OOg and 0*0803 H^O. » 440 ; H - 3-8. 
0*2785 „ 0*2145 Agl. 1-41*6. 

CuHiiI02 requires = 43*7 ; H » 3*6 ; I « 420 per cent. 

SUwr Salt. — A solution of the potassium salt was obtained by 
satorating alcoholic caustic potash solution with the iodolactone, and 
sfter precipitating the excess of lactone by water, silver nitrate was 
added to the filtrate. The silver salt was precipitated as an 
amorphous powder which became crystalline on standing. It is 
•dable in hot water, but cannot be recrystallised without decom- 

0*2430 gave 0*13 ] 5 Ai^L Ag » 24*9 

Oj^jjIOsAg requires Aga 25*3 per cent. 

C«ifiaKno4ac<<ws,CH.'CgH4<^^C>CO.— The iodolactone is dis- 

solTsd in hot alcoholic potash solution and the solution, diluted with 
water, is reduced with 2*5 per cent, sodium amalgam. No hydrogen 
is evolved until almost the theoretical amount of amalgam has been 
added, but it is found advantageous to use a considerable excess of the 
reducing agent in order to ensure the complete removal of iodine. After 
the alcohol has been removed by boiling, the alkaline solution is acidified 
yOL. LXXV. Digitized by^oogle 


by sulphuric acid^ and steam distilled until the distillate no longer 
becomes turbid on adding a strong solution of potassium carbonate. 
The lactone is obtained by salting out the distillate with potassium 
carbonate, and extracting with ether ; the oily residue left on evapo- 
rating the ether distilled almost to the last drop at 290*5^ (unoorr.) 
under a pressure of 768 mm. When purified by redistillation under 
diminished pressure, it was obtained as a colourless, highly refracting 
liquid of sp. gr. at 20^20° -1-0833, and boiling at 126^ (unconr.) 
(20 mm.). The yield of pure cannabino-lactone from the iodolactone 
amounts to 86 per cent, of the theoretical. 

01105 gave 0-3056 00, and 0-0700 HjO. = 74-9 ; H = 7-0. 
CiiH^2^a requires = 75 "0 ; F = 6-8 per cent. 

The lactone dissolves slowly in aqueous solution of caustic alkalis, 
and is reprecipitated by carbon dioxide. So great is the tendency to 
lactone formation that a current of steam slowly removes the lactone 
from its solution in excess of caustic potash. 

Cannahinolact(mieAcid,(XyOK*0f;K^<^^ grams of 

cannabino-lactone, dissolved in hot aqueous potash, was treated with 41 -4 
grams of potassium permanganate dissolved in 1000 c.c. of water, and 
the mixture, after being boiled for 9 hours, was decolorised by a few 
drops of alcohol. The filtrate from the manganese dioxide, when 
evaporated to 200 c.c. and acidified with hydrochloric acid, deposited 
10*6 grams of a pearly-white, crystalline compound, which, after re- 
ci-ystallisation from hot water, was obtained in long needles melting 
at 203^ (uncorr.). The yield is 88 per cent, of the theoretical. Two 
samples were analysed. 

01371 gave 0-3205 00, and 00636 H,0. = 638 ; H = 5-2. 
01380 „ 0-3250 00, „ 0-0626 H,0. = 64-2; H = 50- 
OuHjqO^ requires = 64-1 ; H = 4-9 per cent. 

The lactonic acid is soluble in about 86 parts of boiling water, very 
sparingly in cold water, and easily in alcohol. 

Fbktsnum Salt, — Prepared by adding alcoholic potash to an alcoholic 
solution of the lactonic acid. 

0-2296 gave 00826 K^SO^. K = 16-1. 

Oi^HqO^K requires K = 16-0 per cent. 

The silver salt was found to contain 33-8 per cent. Ag. OjjHgO^Ag 
requires 34*6 per cent. 

The ethylic salt was prepared by boiling the lactonic acid for 6 hours 
with a 3 per cent, solution of hydrogen chloride in absolute alcohol. 
After recrystallisation from dilute alcohol, it melted at 105^ (uncorr.), 
and, on analysis, gave the following numbers. 

Digitized by VjOOQIC 


0-0861 gave 02095 COg and 0-0484 H,0. C = 664 ; H = 6-2. 
01608 „ 0-3905 CO, „ 00890 HgO. C = 66-2; H = 61. 
^n^A- CgHj requires C = 667 ; H = 60 per cent. 

Fusion of Cafmabino-lactonic Acid toith Potash. 

4*4 grains of the lactonic add was fused with 25 grams moist caustic 
potash at 286*^ (in a bath of boiling )9-naphthol). The reaction is 
completed ahnost instantaneously at this temperature, but proceeds 
with extreme slowness at 220^. The melt, on being dissolved and 
acidified with dilute sulphuric acid, gave 3*4 grams of a crystalline 
precipitate, which, after several recrystallisations from dilute alcohol, 
sablimation, and a further recrystallisation, melted above 300°, and 
sublimed without decomposition. 

0-11 33 gave 02392 00^ and 00400 H^O. = 576 ; H = 39. 
OgH^O^ requires = 57*8 ; H = 36 per cent. 

The analysis and physical characters agree with those of isophthalic 

Its methylic salt melted at 64° (unoorr.), and boiled at 280—282° 
(unoorr.) (Baeyer gives 64—65°, Atmalen, 1873, 106, 340). The salt, 
on analysis, gave the following numbers. 

0-1084 gave 0-2445 GO^ and 00518 H^O. = 615; H = 5-3. 
0-1060 „ 0-2395 COj „ 00510 H,0. = 61-7 ; H = 53. 
O^H4(COOOH3), requires = 61-9 ; H =5-1 per cent. 

Potash Fusion of Cann<Mno4actone with Potash, 

The conclusions to be drawn from the fusion of cannabino-lactonic 
acid are supported by the potash fusion of cannabino-lactone (0*9 gram) 
with moist caustic potash (15 grams) at 300° to 320°, which yields 
metatoluic add. 

The reaction took place slowly, and the melt, which was very dark 
in colour, was dissolved, acidified with sulphuric acid, and extracted 
with ether, Ac., in the usual way. After purifying the product by 
ciystaUisation from water and animal charcoal 03 gram was obtained. 
It began to melt at 108°, but was not completely melted until 200°. It 
was accordingly steam distilled, and the crystals, which separated in the 
distillate, were now found to melt at 110° (uncorr.). Jacobsen {Ber.f 
1881, 14, 2349) gives the melting point of metatoluic acid as 110-5°. 
The mother liquor from which the metatoluic acid had been removed 
by steam distillation, on being evaporated to dryness, dried at 150°, 
and then recrystallised from water, gave isophthalic acid, which 
melted above 300°, and sublimed without decomposition ; this was, 
no doubt, produced frdm the metatoluic acid by further oxidation. 

_. „„ p_2oogle 


Reduction qf CannalnnO'laetonie Acid. 

The lactonio acid was reduced by heating with hjdriodic acid and 
phosphorus in a sealed tube at 190% and the product, after two 
recrystallisations, melted at about 210°. As the limited quantity 
of the substance did not allow of further purification, it was analysed. 
The result leaves no doubt that it is the expected metacarbozy- 
phenylbutyrsc acid. 

(1) COOH-C^H^-QgHa-OOOH (3). 

01298 gave 03044 COg and 00696 H^O. = 640 ; H- 59. 
OiiH^O^ requires = 63-6 ; H» 5*8 per cent. 

Uniyersitt Chbmioal Laboratory, 

IV. — Characterisation of JRacemic Compounds. 

By Fbedbbio Stanley Kipping and William Jackson Pops. 

Wb have previously shown (Trans., 1897, 71, 989) that the results of the 
study of a number of externally compensated substances indicate that 
only one method is at present of really practical use for characterising 
solid racemio compounds, that, namely, which is based on the deter- 
mination of the crystalline forms of the optically active and externally 
compensated materials ; and, in accordance with this conclusion, we 
defined crystalline racemic and pseudoracemio compounds from a purely 
crystallographic standpoint {loe. eit., 993). But the crystallographic 
constants of an organic substance are not always determinable with 
ease and completeness ; hence the establishment of other criteria of 
racemism is a matter of considerable importance. 

Now, liebisch has shown (AnnaieHf 1895, 286, 140) that a com- 
parison of the densities of the optically active and externally compen- 
sated compounds affords a very simple method for the determination of 
racemism (compare Walden, B&r.^ 1896, 29, 1692); unfortunately, 
however, the results of the experimental determination of a difference 
between two density constants, especially when this difference is very 
small, as it frequently may be, can hardly be considered so conclusive 
as those derived from the much more complex series of constants 
constituting the crystallographic properties. 

This difficulty of ascertaining, except by crystallographic deter- 
minations, whether certain substances are really racemic on the one 
hand, or merely pseudoracemio or externally compensated mixtures on 

Digitized by VjOOQIC 


the other, baa led to attempts being made to devise simpler methods for 
the determination of racemism. Thus, Ladenburg {Ber„ 1894, 27, 
3065) formolates " eine allgemeine Methode^ am Qemenge enantio- 
morpher Korper von racemischen Yerbindongen zu unterscheiden. 
Sobald es gelingt, die in Frage stehende Snbstanz, die wenigstens 
einen kleinen Ueberschass der einen drehenden Modification enthalten 
moss, dnrch Behandlnng mit inactiren Korpem in Fractionen yon 
▼eriLndertem Drehnngsvermogen zu Terwandeln, liegt stets eine 
Yerbindnng Tor." 

Now this method, if really applicable, should prove of great value, 
but inasmuch as Ladenburg adduces no experimental evidence sup- 
porting its validity as a means of discrimination, but uses it without 
further inquiry in order to decide as to the racemic nature of 
externally compensated coniine, the conclusions based on its use cannot 
be regarded as in any way decisive. 

Moreover, the method is arrived at by means of a curiously 
falladous piece of reasoning, for, just before the above quotation, 
Ladenbuzg writes : ^ es ist, wie ich glaube, eine bisher ausnahmlos 
bestiltigte Thatsache, dass enantiomorphe KQrper stets die gleiche 
Lbdiehkeit besitaen und daher auch in Yerbindung mit inactiven 
Stoffen nicht duroh Erystallisation oder partielle FalluDg getrennt 
werden konnen." 

Kow there seems not the least justification, experimental or 
theoretical, for the deduction which Ladenburg draws from the 
fact of the equal solubility of the two enantiomorphously related 
substances, namely, that a mixture of unequal quantities could not be 
separated by crystallisation or partial precipitation. There is, in fact, 
no evidence that an inactive non-racemic mixture has the same 
solubility as either of its active components and if its solubility be 
greater, then, on concentrating a solution containing such a non- 
raoemic mixture together with excess of one enantiomorph, at constant 
temperature, that excess would crystallise first, leaving material of 
lower specific rotation in solution ; a continuation of this process would 
ultimately afford a mother liquor containing an inactive non-racemic 
mixture. The solution would then go on depositing dextro- and l»vo- 
material in equal proportion as evaporation proceeded, casual dis- 
turbanoes of equilibrium such as always occur in a crystallising 
solution being, of course, disregarded. Obviously, therefore, any 
arguments based upon our present knowledge of the laws governing 
soIubiUty should have led Ladenburg to a conclusion diametrically 
opposed to that at which he actually arrived. 

The question of the racemic nature of externally compensated coniine 
has lately been again considered, and Kiister {Ber., 1898, 31, 1847), 
arguing from tj|te solubility products of the various isomerides, con- 

Digitized by VjOOQIC 


eludes that a large part of the inactive bubstance exists in solution in a 
raoemic condition. Although this conclusion may be true, yet, 
inasmuch as no case has been investigated in which the solubUity 
values indicate that no raoemism occurs in the solution, or in the solid 
state, as the case may be, Kiister's method cannot be accepted. A 
hypothesis founded upon one set of phenomena often gives valuable 
indications as to the direction in which work should be done in con- 
nection with the examination of a second set of phenomena, but these 
are merely indications, and require experimental verification before the 
hypothesis can be extended so as to cover the new ground. The cases 
dealt with by Kiister are instances in which the ordinary solubility 
laws applying to two mutually inactive solutes are not followed ; when 
a case is found amongst externally compensated substances in which 
the solubility determinations indicate the non-existenoe of a racemic 
compound in solution, Kiister's method will become of practical value. 
Until then, however, it must be classed with Ladenburg's method as an 
untried one. 

In order to ascertain the accuracy or otherwise of Ladenburg's 
argument, we have examined three cases experimentally; those, 
namely, of sodium ammonium tartrate, sodium potassium tartrate, 
and potassium hydrogen tartrate. 

Sodium Ammonium Dextro- and LcBvo-tartraiea, 

An externally compensated mixture of dextro- and levo-sodium 
ammonium tartrates is known to be non-racemic at ordinary tempera- 
tures, but becomes racemic at and above 27^ (van't HofE and 
Deventer, £er,, 1886, 19, 2148); below this temperature, the salt 
separates from solution as a mere mixture of the two tartrates. 

Now, if Ladenburg's rule be applicable, on crystallising the sodium 
ammonium salt of racemic acid with an excess of sodium ammonium 
dextrotartrate below 27^> the successive fractions should have the 
same specific rotation ; but, as a matter of fact, the behaviour of the 
mixture on crystallisation is that which would be expected from our 
interpretation of the laws governing the equilibrium of such solu- 

In a preliminary experiment, an intimate mixtture of 25 grams 
of sodium ammonium dextrotartrate with 5 grams of the corres- 
ponding lisvotartrate was made ; it was found to have the specific 
rotation [a]i>e +15-60^ in a 5 per cent, solution, instead of the 
calculated value [a] « + 16*76^ The mixture was dissolved in cold 
water and set aside to evaporate; after several days, a fraction (1) aepa- 
rated and was collected, washed with cold water, and dried in the air. 
The mother liquor and washings were mixed and set aside to crystallise^ 

„ .gitized by VjOOQ _ _ 


when a further separation (2) was obtained ; this was removed as 
before, the mother liquors evaporated to dryness, and the residue 
collected. During the experiment, the laboratory temperature never 
rose above 15° ; the weights and rotations of these fractions were as 

Weight [a]D 

Total material 30 g. + IS-eO"" 

First fraction, (1) 8 g, +23*51 

Second „ (2) 13 g. +20-27 

Third „ (3) 8 g. 

Instead of- all the three separations having the same specific 
rotation as the original mixture, as required by Ladenburg's rule, the 
first two fractions consisted almost entirely of dextrotartrate, whilst 
a 5 per cent, aqueous solution of the residue in the mother liquor had 
no observable rotation in a 200 mm. tube. 

In order to trace this process of separation more carefully, and thus 
obtain further data, the following experiments were made. A mixture 
of 26*1 grams of sodium ammonium dextrotartrate, with an equal 
weight of the salt obtained from racemic acid, wiis made ; this mixture, 
having a specific rotation of [a]o= + 11*82°, was dissolved in water 
and allowed to crystallise spontaneously at the ordinary laboratory 
temperature, which never rose above 18°, namely, 9° below the tem- 
perature at which a racemic compound begins to separate. As the 
various deposits were obtained, they were separated by means of the 
filter pump (but not washed free from mother liquor) and examined, 
the fractionation proceeding in accordance with the following scheme 
(next page). 

The weights, w, and the specific rotations a, of the various fractions 
are given in columns 2 and 3 respectively of Table I (p. 41), columns 
5 and 6 give respectively the excess of dextro- over inactive salt in 100 
parts of the various fractions, and the actual weights in grams of the 
excess of dextro-salt in these fractions. 

The weight of salt recovered from the 52*2 grams dissolved was 2to 
»51-25 grams; the weighted mean specific rotation of all the 
fractions is 2u^/2io» +12-20°, and if the weight of salt taken, 
namely, 62*2 grams be substituted for 2to, the mean specific rotation 
2ioa/52*2= +11*98° is obtained. 

The mixture of salts actually had the specific rotation [ajo « + 1 1*82° 
so that within the limits of the experimental errors unavoidably 
incurred in so long a series of fractionations, the whole of the salt used 
is accounted for. 

The specific rotation of sodium ammonium dextrotartrate, 
Na(NH,)C,H,Oe + 4H20i 
is [ajo^ +23*64°, and the first large deposit of 32*1 grams contaiued 

„.gitized by Google 





d rf 










— to 






« i 






^ tCQ 











Digitized by VjOOQIC 

Table I. 


1. 2. 











excess of 


Weight of 

excess of 








+ 28-28 
+ 28-28 
+ 21*66 

+ 28-47 
+ 28-12 
+ 28-02 

+ 22-61 


+ 16*62 

+ 6-41 

+ 112-67 
+ 76-60 
+ 86*64 

+ 86*20 

+ 106-86 

+ 89-78 

+ 66-27 

+ 64-11 

+ 64-41 

+ 98-26 
+ 98*26 
+ 90-86 

+ 99*28 
+ 97-80 
+ 97*88 

+ 96*22 


+ 66*66 

+ 22-88 

+ 4-766 
+ 8194 
+ 1*660 

+ 1-488 
+ 4-499 
+ 8-798 

+ 2-881 

+ 2-724 
+ 2-666 


2u«= +626-29. 

3IP= +26-461. 

????= +12-20'. 

|^= +11-98'* 

28*25 grams of sodium ammonium deztrotartrate and 3*85 grams of 
the isomeric levo-salt ; this corresponds with 24*4 grams of the sodium 
ammonium deztrotartrate with 7*7 gramsof the inactive mizturOi so that 
nearly the whole of the 26*1 grams of the dextro-salt originally 
taken was deposited in the first crop of crystals. After the next 
separation, m^, from the mother liquor, 27*1 grams of the deztrotar- 
trate had separated in addition to inactive material ; consequently, the 
mother liquor was strongly Isevorotatory, and the next separation, m,, 
was a ]«vorotatory one. Practically, all the excess of the dextro- 
rotatory salt was contained in the first fractions, A^, h^, Ag, A^, k^ k^ and 
l^f which consisted almost entirely of this salt, but still contained 
small quantities of the inactive material, partly because the crystals 
were not washed, and partly, no doubt, because of occluded mother 
liquor, which is often present in noticeable quantity. The further 
flanges during crystallisation will be understood from the tabulated 

In a recent paper (Trans., 1897, 71, 999), we have shown that ex- 
temally compensated camphorsulphonic chloride (Trans., 1893,63^ 560) 
and camphorsulphonic bromide (Trans., 1895, 07, 359) are probably 
peeudoraoemic ; consequently, they behave, on crystallisation, like non- 

Digitized by VjOOQIC 


racemic compounds and the ethylic acetate solutions of a strongly 
dextrorotatory mixture of the sulphonic chlorides deposits^on evapora- 
tion, part, or the whole, of the excess of the dextro-compound ; after 
the solution has thus become nearly or quite inactive, further crystal- 
lisation affords deposits which sometimes contain an excess of one or 
other isomeride, a behaviour which is obviously very similar to that of 
the sodium ammonium tartrates. 

JSodium Fotaasium DextrotartraU and Racemaie, 

Having shown by the foregoing experiments that Ladenburg'a 
method does not hold in the case of non-racemic sodium ammonium 
dextro- and Isavo-tartrates, namely, in the only test case on which 
the method has yet been worked out, we thought it advisable to 
apply the same method to a case in which a crystalline racemic com- 
pound undoubtedly exists, in order to ascertain whether the fractional 
separation occurs in a manner essentially different from that observed 
in the case of a non-racemic mixture. 

Sodium potassium dextrotartrate is isomorphous with the^ corre- 
sponding sodium ammonium salt and, at temperatures between — 6^ 
and + 41% forms a racemic compound with the isomeric Invotartrate 
(van't Hoff and Deventer, Zeit. phyaih. Chem,, 1895, 17, 505)3 it 
therefore forms a racemic compound at ordinary temperatures. The 
optically active substances have the composition NaKO^H^O^ + 4H2O, 
whilst the composition of the racemate is NaEO^H^O^ + SH^O. 

A mixture of 47*4 grams of sodium potassium dextrotartrate with 
44*4 grams of the racemate (molecular proportions) was dissolved 
in water and fractionally crystallised as before ; the scheme given 
on the next page shows how the various fractions were collected. 
Table II (p. 44) gives the weights, to, and the specific rotations, a, for 
the Z>-line ; columns 5 and 6 give the values corresponding with those 
in Table I. 

The specific rotation of the. anhydrous dextrotartrate, KNaO^H^O^, 
is [ajo-" +29*67° (Landolt) ; that of the mixture of 47*4 grams of 
hydrated dextrotartrate with 44*4 grams of the racemic salt is, 
therefore, [a]]) » +11*41°. The weighted mean specific rotation of 
all the separated fractions is 


?^= +11*28°. 

Digitized by VjOOQIC 







«■ J? 

c0 m 




h4 ** 





V to 

Digitized by VjOOQIC 


Tablb II. 












excess of 


Weight of 

excess of 




+ 22-08" 

+ 202 

+ 99-78 




+ 21-77 

+ 226-6 

+ 98*88 

+ 10*280 



+ 21-66 

+ 160*6 

+ 97*88 

+ 7-266 



+ 21-26 


+ 96*98 

+ 4-271 



+ 17-60 

+ 162-2 

+ 79-22 

+ 6-891 



+ 21-86 

+ 166-6 

+ 96*48 

+ 7*628 



+ 4-68 

+ 66-9 

+ 81-90 

+ 8-864 









+ 10-00 

+ 84-0 

+ 46-17 

+ 1*636 



+ 1-86 

+ 9-0 

+ 6-10 

• +1*408 






2wo= +1086-4. 


SZ>= +47*622 



= +11*28'. 

The agreement between the quantities and specific rotations of the 
material taken, and the quantities and specific rotations of the 
fractions recovered, is hence highly satisfactory. 

From an inspection of Table II, it is seen that the successive 
fractions separating from the solution decrease in specific rotation, 
that is to say, they contain decreasing quantities of the deztrotartrate 
in eiccess of the racemate, and ultimately the mother liquors contain 
nothing but pure racemate. The table also shows the following 
curious facts : the fractions h, k, and I contain 48*66 grams of the 
deztrotartrate as such, whilst material corresponding with only 47*4 
grams was used; the difference of 1*26 grams, therefore, must be 
ascribed either to experimental error or to the fact that part of the 
racemate used was resolved, and the dextrorotatory component de- 
posited here. That the latter alternative is the true one is shown by 
the mother liquor becoming Isovorotatory. Consequently, on crystal- 
lising the latter, the first deposit, m^, is strongly IsBvorotatory, con- 
taining 90*34 grams of l»votartrate to each 9*66 grams of racemate; 
2*98 grams of l»votartrate crystallised as such in m^, whilst the 
solution only contained 1*26 grams excess of this salt. In the end, 
therefore, 50*38 (47*4 + 2-98) grams of dextrotartrate had separated 

Digitized by VjOOQIC 


from solution as such, or about 3 grams more than had been used ; 
it follows, therefore, that a partial resolution of the racemate had 
occurred, about 6 grams being so resolved. The good agreement of 
the sums of the weights of salt separated from the solution and the 
mean rotation of the fractions, on the one hand, with the weight of 
salt and the mean specific rotation which should have been obt^ned, 
on the other hand, is proof that we are not here being misled by 
experimental error. 

The interesting work of Purdie on the resolution of lactic acid into 
its optically active components (Trans., 1893, 63, 1143) affords a 
case, similar to the above, of the resolution of a racemic compound. 
Baoemic zinc ammonium lactate crystallises with SH^O, and when its 
supersaturated solution is sown with a crystal of either of the 
optically active isomerides, that particular isomeride separates in 
crystals containing 2H2O ; this resolution is similar to that which 
has occurred in fractions l^ and m^ of our mixture. 

We are thus led to the following conclusion : a racemic compound 
may, under certain conditions, be resolved into its optically active 
components by simple crystallisation, at temperatures at which the 
racemic compound is more stable than the mixture of the two optically 
active salts. 

A comparison of Tables I and II shows that the fractional crystal- 
Uaation has followed much the same course in the case of the racemic 
compound as in that of the non-racemic mixture ; the separation of 
the racemic compound from the dextrorotatory one is, however, not 
quite so sharp as in the separation of the non-racemio mixture. 

Poiasnvm Hydrogen Dexiroiartrctte and Racemate^ KHC^H^O^. 

In the foregoing cases, the inactive mixture or compound is more 
soluble than either of the active components; it seemed desirable, 
therefore, to investigate a case in which the contrary is true. Such 
a ease was found on examining the behaviour of the potassium hydro- 
gen tartrates. On crystallising a mixture of equal quantities of the 
deztrotartrate and of the racemate, by allowing the hot solution to 
cool, it was found that the mother liquor contained a salt of much 
higher specific rotation than that of the mixture deposited; by 
repeatedly recrystallising the deposit, the specific rotation of the 
material in the successive mother liquors remained nearly constant, 
and of very much higher value than that of the crystalline separations. 
Owing to the sparing solubility of these two salts, they had to be dis- 
solved in dilute ammonia for the rotation determinationa The results 
need not be quoted in full, because they were in so many respects similar 
to the preceding cases, and the experimental error was greater. 

Digitized by VjOOQIC 


46 pope: crystalline form of iodoform. 

The ezanunation of these three cases shows that Ladenburg's method 
does not constitute a means of discriminating between cases in which 
a solid racemic compound is formed, and those in which a more inactive 
mixture is obtained ; a scrutiny of Tables I and II will show that this 
method gives practically the same results with a wellniefined racemic 
compound as with a non-racemie mixture. 

It may possibly be objected, that, although inactive sodium am- 
monium tartrate does not exist as a solid racemic compound at 
ordinary temperatures, it may exist as a racemic compound in solu- 
tion ; to make this assumption unsupported by experimental evidence 
is, however, unjustifiable, and can only be regarded as an expression of 

Chemistby Depa&tment, Chemistbt Department, 

Uniybrsity College, Goldsmiths' iNSTiruTE, 

Nottingham. London, S.E. 

V. — Crystalline Form of Iodoform. 

By William Jackson Pope. 

Up to the present, no geometrical measurements with any pretensions 
to accuracy have been made on crystals of iodoform, the only available 
data consisting of two angles measured by Bammelsberg {Krytt.-pkyM, 
Chem., 1882,2, 321) ; these measurements cannot, however, be credited 
with much value since Eammelsberg remarks of his crystals that *' die 
Flachen sind ziemlich matt." 

Considerable difficulty is found in obtaining well developed crystals of 
iodoform using the ordinary organic solvents such as alcohol or benzene. 
Iodoform is, however, fairly soluble in acetone, and on allowing 
the cold solution to evaporate spontaneously at a uniform temperature, 
magnificent, transparent, six-sided tablets of iodoform separate ; in order 
to obtain the best results, the acetone should be as free from water as 
possible. Even when pure, almost odourless iodoform is used, blacken- 
ing occurs round the sides of the vessel, and some product, probably 
iodacetone, is formed during the evaporation which excites to tears, 
and powerfully affects the mucous membrane. 

The crystals, which may readily be obtained of a centimetre in dia- 
meter, and several millimetres in thickness, are uniformly developed, 
and the faces give very perfect reflections on measurement. The 
dominant form is c{lll}, the pyramid r{100, 22T} being very much 
smaller ; the prisms p {TtO} and m {211} are rarely observed and are 
always small (see Fig. 1). The crystals are very hard and brittle, free 

„.gitized by Google 

pope: crystalline FaRM OF IODOFORM. 4*7 

from siriations, and have no noticeable cleavage ; they show a six-sided 
internal growth of hexagonal symmetry when allowed to grow rapidly. 

The normal hexagonal optic axial interference figure is seen on cono- 
seopic examination through the faees of o {111} ; the double refraction 
is strong and negative in sign. 

Bammelsberg (Joe. cit,) determined the angles cr (111 : 100)^52'' 0' 
and rr (22T : 100) ^iG"" 30', gives the axial ratio a : c»- 1 : M08, and 
only observed the forms e{lll} and r {100, 221}. 

Crystalline system. — ^Hexagonal. 

a:c = 1 : 1-1084. a- 93^41'. 

Fonns observed c {111}....^ {0001} 

r {100, 221} {1011} 

pjllO} {2110} 

w{2TT} {OlIO} 

Fig. 1. 


Tbe f oUowing angular measurements were obtained. 


Knmber of 






51*55'— 52° r 





85 68—88 6 

86 2 

86° 2'20'' 



75 50— 78 6 

76 30 

76 20 



98 54— 94 8 

98 58 10 

98 57 40 



46 43—47 2 

46 57 80 

46 58 50 



46 20 — 46 25 

46 23 40 

46 24 



133 82 —183 41 

138 36 10 

138 36 



89 67—90 4 

90 20 




87 54—88 9 

88 2 

88 010 

Ajftercantiousmeltuig under acover slip on a microscopeslide, iodoform 
erystallisee readily, giving broad, individual crystal flakes, the surfaces 
of which are parallel to e {111} ; the optic axis of negative double re- 
fraction emerges normally to the film surface. 

li is interesting to note that a large proportion of compounds of 
simple oonstitution crystallise in systems of high symmetry, such as 
the cubic or hexagonal. It may also be remarked that, whilst the 
molecule of carbon tetriodide has partial cubic symmetry and crystaf- 
lises in the cubic system, the molecule of iodoform has partial hexagonal 

Digitized by VjOOQ IC 


symmetry and crystaUises in the hexagonal system. Many other 
similar analogies could be quoted, and although complications due to 
polymorphism frequently occur, it may be stated that a knowledge of 
the chemical constitutions of simple compounds often allows of a safe 
prognosis of the crystalline form which those compounds assume. 

Ghbmioal Depabtmknt, 

CsMTaAL Tbohnioal College, London. 

VI. — fifi'Dimethylglutaric Acid and its Derivatives; 
Synthesis of cis- and trans- C7aronic Adds. 

By William H. Perkin, jun., and Jocbltn F. Thorpe. 

Cabonb, OioHj^jO, one of the most important ring ketones in the 
terpene series, is formed when dihydrocarvone hydrobromide is treated 
with alcoholic potash, hydrogen bromide being eliminated, a decompo- 
sition which, according to G. Wagner,* may be formulated in the 
following manner. 

CH-CHj • CH-CHg 

H2C CO / 


HjC CH2 = I (CHg), I + HBr. 

Sy HjC C CH 

(CH3)26Br \^^ 

Dihydrocanrone hydrobromide. Carone. 

This view of the constitution of carone was considered probable by 
Baeyer, who, in order to confirm this formula, carried out a number 
of important experiments on carone, during the course of which he, 
in conjunction with Ipatieff, investigated the behaviour of this sub- 
stance on oxidation with permanganate {Bw., 1896, 29, 2796). 
It was found that, although carone is very stable towards perman- 
ganate at the ordinary temperature, it is moderately readily oxidised 
at 100^, with formation of two isomeric dibasic acids, C5H3(COOH)2y 
melting at 176^ and 212^, which were named caronic acids. Tlxe 
former of these, which is produced in much the larger quantity, 
readily yielded an anhydride melting at 54 — 56° when boiled with 
acetyl chloride, but the other acid was not affected by this treatment. 

* Compare Baeyer {Ber,, 1896, 29, 5 and 2796). 

Digitized by VjOOQIC 


A careful examination of these acids led Baeyer and Ipatieff to the 
conclusion that the caronic acids were stereoisomeric modifications of 
dimethyltrimethylenedicarboxylic acid. 

0(CH3), C(CH,), 

/ /\ 



fndu-Cuoiiic acid. ei»-Cuoiiio acid. 

and that these were formed by the ozidstion of carone at the points 
indicated by the dotted lines in the formula 


H.C C — CH 



That these acids have the same structure and are stereoisomeric 
seemed probable from their behaviour towards hydrobromic acid at 
100^, under which conditions both are converted into terebic acid with 
disruption of the trimethylene ring {loc. cU,, p. 2801). 





'(^^8)2 + HBr. 


Ho mention is, however, made of any attempt to convert the one 
modL€cation into the other. 

In studying this important work, it seemed to us that it would be 
most interesting to find some means of synthesising the caronic acids, 
and of thus placing their constitution beyond doubt. This was 
ultimately accomplished in the way described in this paper. 

Some time since, it was shown by Goodwin and Ferkin (Trans., 
1896, 60, 1476), that ethylic dimethylaciylate condenses with the 
sodium derivative of ethylic malonate with formation of ethylic 

{OOOCJB^^)fin' C(CH8)2' CHj- COOC2H5, 
and from this, by hydrolysis and elimination of carbon dioxide, )3^-di- 
methylglutaric acid, COOH- CH^- C(CH3)2- CH^* COOH, was prepared. 
The yield of acid obtained in this way is small, but by substi- 

Digitized by VjOOQIC 

50 PfiRKlK AND TltORPfe : /8/8-t)IMETHYLatUTARtC 

tilting ethylic cyaaacetate for ethylic malonate in the condensation 
with, ethy lie dimethylacrylate, wenonrfind that the yield of condensation 
product, which consists of a mixture of some ethylic a-cyaruhfiP-difrntkyl- 
glutarate, COOCgHj- CH(CN)-C(CHg)j- CHg- COOCgHg, with much of the 
hydrogen ethylic salt, COOH-CH(CN)-C(CHs)2-CHj-COOC2H6' " 
more than 80 per cent, of the theoretical, and, as these ethereal 
salts, on boiling with 50 per cent, sulphuric acid, are quantitatively 
converted into ^^-dimethylglutaric acid, it is now an easy matter to 
prepare this acid in quantity. 

When the anhydride of dimethylglataric acid is treated with phos- 
phorus pentabromide and bromine, and the product poured into absolute 
alcohol, eifiylic bramodimethylgltUawate, 

COOCaHg- CHBr • C(CH3)2-CH2- COOCgH^, 
is produced, together with large quantities of the hydrogen ethylic 
salt of the same acid, COOH- CHBr •C(CH3)2-CH2-COOC2H5 ; that the 
latter should be produced in such quantities is certainly remarkable, 
and a possible explanation of its formation is given in the experi- 
mental part of this paper. When ethylic bromodimethylglutarate is 
digested with alcoholic potash, hydrolysis and elimination of hydrogen 
bromide takes place simultaneously, and a mixture of acids is ob- 
tained which, by conversion into the ammonium salts and treatment 
with alcohol, as recommended by Baeyer and Ipatieff {loc, dt,, 2978), 
is easily separated, and found to consist of large quantities of trans- 
ca/ronic addj some lactone acid of hydroxydimethylglutaric acid (see 
below), and traces of cis-caromc acid. 

This synthesis of the caronic acids may be represented in the 
following way. 

C(CH3), C(CH,), 

/ \ gives /\ 


Hydrogen ethylic bromodimethylglutarate, 

on treatment with alcoholic potash, is quantitatively converted into 
^rotw-caronic acid, apparently without even traces of the cts-modifica- 
tion being formed. The synthesis of a trimethylene compound in the 
manner represented above is remarkable, but a few similar cases have 
been observed, as, for example, in the formation of acetyltrimethylene 
by the action of alkalis on acetylpropyl bromide, 

CHa'CO-CHg^^^ « CH3-CO-CH<5^2 + HBr. 

The synthetical caronic acids have been very carefully investigated, 
and it is shown in the experimental part of this paper that their pro- 
Digitized by VjOOQ l€ 


periiee agree in all respects with those of .the acids obtained by Baeyer 
and Ipatieff from carone, so that there cannot be any doubt as to tha 
identity of the acids produced by these different methods. 

One additional point of interest has been discovered, namely, that 
IratMHsaionic acid is converted into the anhydride of cM-caronic add 
by the action of acetic anhydride at 220^, a transformation of the 
fiww-into the ctt-modification which was required to clearly show 
that these two acids are stereoisomeric. 

Furthermore, it is shown in this paper that aqueous sodium car- 
bonate hydrolyses the hydrogen ethylic salt of bromodimethylglutaric 
acid, CXX)H* CRBt*0{OH^)2* CH^* COOO^H^, in a manner quite different 
from alcoholic potash, with formation of the IctoUme of a-hydroxy- 

talline substance which melts at 112*^ and is isomeric with the 
caronic adds. 

When ethylic bromodimethylglutarate, 

0000^5- CHBr- C(CH8)j- CHj-OOOOjHj, 
is digested with diethylaniline, hydrogen bromide and ethylic bromide 
are eliminated and an ethereal salt is obtained, which, on examination, 
has been found to consist of the ethylic salts of ^arM-caronic add, and 
of the laeUmiB of hydroxydimethylglutaric acid. 


CondenMion of Ethylic JHmeihylacrylaie toiih tiie Sodium Derivative qf 
Ethylic CyanaceUOe. 

^)8.Dimethylglutario acid, OOOH-OHj-C(CH8)j-CH3- COOH, the acid 
which is the starting-point in the synthesis of the caronic adds, was 
fint prepared by Goodwin and Ferkin (Trans., 1896, 69, 1475), who 
obtained it by the following series of reactions. 

Ethylic dimethylaerylate was^ in the first place, digested with the 
sodiom derivative of ethyUo malonate in alcoholic solution, when con- 
densation took place with formation of ethylic dimethylpropanetri- 

(00O0jH5)2OH2 + (0H3)^0:CH- COOCjHg = 

(COOO,H^)jCH- C(CH3)j- CHa'COOOjH^. 

This ethereal salt, on hydrolysis, yields the corresponding tribasie 
add, which, at 200^, loses carbon dioxide with formation of )3^-dimethyl- 
glataric add, OOOH*CH2-C(CH3)2-CHj-OOOH, and from this the 
anhydride is readily prepared by treatment with acetic anhydride. 
The yield of the original triethylic salt is unfortunately not good, and 

Digitizedfy Google 


seldom reaches more than 40 per cent, of the theoretical ; indeed, the 
average yield is scarcely more than 28 per cent. Until quite recently, 
* we have prepared all the dimethylglutaric anhydride required for this 
research by the method devised by Goodwin and Perkin. A few months 
since, however, we discovered that the yield may be very greatly im- 
proved by substituting ethylic cyanacetate for ethylic malonate in the 
condensation with ethylic dimethylacrylate, the yield of condensation 
product being increased to at least 80 per cent, by this means.* 
Ethylic a-cyano-pp-dimethylgliUarcUe, 

COOCjHg- CH(ON)-C(CH8)2-CH2- COOCgHg, 
has been prepared in large quantities by the above process, the 
details of preparation being the following. Sodium (23 grams) is 
dissolved in alcohol (300 grams), the solution of sodium ethoxide 
mixed with ethylic cyanacetate (113 grams), ethylic dimethyl- 
acrylate (129 grams) is then added, and the whole heated in a reflux 
apparatus on the water bath. The sodium derivative of ethylic cyano- 
acetate, which at first separates as a white, crystalline powder, slowly 
dissolves, and the liquid darkens and gradually sets to an almost solid 
cake of the sodium derivative of the condensation product, the reaction 
being finished in about 24 hours. Water is now added, and the oily con- 
densation product extracted with ether in the usual way; the ethereal 
solution, after washing with water and drying over calcium chloride, 
deposits a thick oil which, after two distillations under reduced pres- 
sure, passes over constantly at 190° (30 mm.), and consists of pure 
ethylic cyanodimethylglutarate. 

0*2614 gave 13-5 c.c. nitrogen at 20"^ and 750 mm. N » 5*80. 
C12H19NO4 requires N = 5-81 per cent. 

The yield of this substance produced in the above reaction is about 
15 per cent, of the theoretical. 

The principal product formed in this condensation is the acid ethylic 
salt of the substance just mentioned, that is, ethylic hydrogen a-cyano-PP* 
difnethylghacuxae, COOH-CH(CN)-C(OH3),-OHj-COOC2H5 ; this is 
obtained on acidifying the mother liquors of the condensation product 
mentioned above and extracting with ether. After thoroughly 
washing the^ethereal solution^ drying over calcium chloride and evapo* 
rating the ether, a thick oil ib left, which was not analysed, since, 
for reasons stated below, it cannot be purified by distillation, and it 
showed no signs of crystallising. The yield of this substance formed 
is no less than 60 — 70 per cent, of the theoretical. 

* This immensely increaaed yield is not confined to this condensation, since i1 
h&a been found that a similar resalt is obtained when other unsaturated etherea] 
salts, such as ethylic crotonate, ethylic methylacrylate, kc, are employed, and tb4 
products formed in this way are at present being investigated by one of us. 


It is very difficult to onderstaDd why this acid ethereal salt should 
be produced in such large quantities ; its formation is certainly not 
due to the presence of water, as several caref al experiments which 
were made with specially dried alcohol gave, in every case, the same 

When distilled under ordinary pressures, this acid ethereal salt is 
readily decomposed, with elimination of carhon dioxide and formation of 
ea^ y-cyano-PP-dtmeihylbutyraU, CN-CH5-C(CH8)3-CH2-COOC2H5, 
which is a mobile oil distilling without decomposition at 244°. 

0-2783 gave 21-8 c.c. of nitrogen at 20° and 735 mm. N = 8-66. 
CjjHjsOgN requires N = 8'28 per cent. 

When digested with concentrated hydrochloric acid in a reflux appa- 
ratas, it rapidly dissolved, and on cooling and mixing with an equal 
bulk of water, a mass of crystals separated, which, on examination, 
was found to consist of the imide of dimethylglutaric acid. 

PP-Dimethylglutartmide, {CK^)^0<!^\^^1!^R. 

This substance is produced by the hydrolysis either of ethylic 
a*cyanodimethylglutarate or hydrogen ethylic cyanodimethylglutarate, 
the process in the two cases being somewhat differently conducted. 
When the first or neutral ethylic salt is employed, the pure substance 
(60 grams) is heated with methyl alcoholic potash (50 grams) for 
2 hours, the solution evaporated until free from alcohol, acidified, and 
extracted with ether. After evaporating the ether, the residue, which 
probably consists principally of cyanodimethylglutaric acid, is digested 
in a reflux apparatus with concentrated hydrochloric acid for 3 hours, 
when, on evaporating, the whole becomes filled with colourless needles 
of the imide ; these are collected with the aid of the pump and recrys- 
tallised from water. In the case of the hydrogen ethylic salt, it is 
only necessary to boil with an equal volume of concentrated hydro- 
chloric acid for 3 hours, in order to get directly an almost quantita- 
tive yield of the imide, care being taken to so moderate the reaction 
as to avoid loss from the evolution of carbonic anhydride, which is apt 
to become very vigorous. 

A^Dimethylglutarimide crystallises from water in long, colourless 
needles, which melt at 144°, and at a higher temperature distil 
unchanged. It is very sparingly soluble in cold water, readily in hot 
water, and is almost insoluble in dry ether. 

0*3034 gave 26 c.c. nitrogen at 17° and 750 mm. N = 9 *82. 
CyH^jNO^ requires N « 9*93 per cent. 

Digitized by VjOOQIC 


This imide is quantitativelj converted into )3)3-dimethylglataric 
acid (m. p. 101^) on heating in a dosed tube with concentrated 
hydrochloric acid for 5 hours at 200^ or by heating with dilute 
sulphuric acid (50 per cent.) for 3 hours on a sand bath, the acid 
being readily obtained from the products of hydrolysis by extraction 
with ether in the usual way. 

Pp-DiTnethylghdaric Anhydride, V^^* ^(CHg)^- CH2 

This anhydride has already been obtained from the acid by treatment 
with acetic anhydride (Trans.; 1896, 69, 1475), but the process used in 
the production of the acid and conversion of this into the anhydride has 
subsequently been much improved, and very large quantities have 
been prepared for the purposes of this research by the following 
method. Ethylic cyanodimethylglutarate, as well as the acid ethylic 
salt, which is always the chief product in the condensation of ethylic 
cyanoacetate with ethylic dimethylacrylate, are boiled with an equal 
volume of 50 per cent, sulphuric acid for 12 hours on the sand bath. 
The exact end point of the reaction is difficult to determine, owing to the 
fact that the sulphuric acid converts a large proportion of the dimethyl- 
glutaric acid into anhydride, and this,' like the unhydrolysed ethylic 
salts, floats on the surface of the aqueous liquid as on oil ; we found, 
however, that 12 hours was sufficient to ensure complete hydrolysis. 

The product, when cold, is extracted several times with ether, the 
ether distilled off, and the residue, without further purification, mixed 
in the same flask with an equal bulk of acetic anhydride, and boiled 
for 3 hours on a sand bath, using a reflux condenser. After distilling 
off most of the acetic anhydride under the ordinary pressure, the 
crude dimethylglutaric anhydride which remains is purified hy 
fractionation under reduced pressure, when almost the whole passes 
over at 181^ (25 mm.) as a colourless oil, which, on cooling, solidifies to 
a hard, crystalline cake. 

The yield of pure anhydride obtained in this way was over 90 per 
cent, of theory calculated from the ethylic cyanodimethylglutarate, or 
about 73 per cent, calculated from the ethylic cyanoacetate originally 
used in the condensation. 

Actum qf Bromine on DimetJiylglutaHc Anhydride, — In investigating 
this reaction, dimethylglutaric anhydride was treated with phosphorus 
pentabromide and bromine, and the bromo- or dibromo-acid bromides 
thus produced were converted into methylic or ethylic salts by the 
action of methylic or ethylic alcohols. 

EHhylie a-bromo-fifi-dimethylgluiaraie, 

COOCgH^- CHBr •C(CH3)j- QH^* COOCaH^, 

Digitized by VjOOQIC 


was prepared from dimethylglutaric anhydride (13 grams) by mixing 
it with phosphorus pentabromide (50 grams), and heating the mizture 
in a refloz apparatus on the water bath until the reaction was com- 
plete ; bromine (16 grams) was then gradually added, and as soon as 
the vigorons evolution of hydrogen bromide had ceased the whole was 
heated on the water bath until colourless, and the product poured into 
well-cooled absolute alcohol. After being allowed to stand for some 
hours, water was added, when a heavy oil was precipitated which was 
extracted with ether, and the ethereal solution, after washing well 
with sodium carbonate solution, was dried over calcium chloride and 
evaporated. The nearly colourless heavy oil thus obtained, on fractiona- 
tion, distilled constantly at 181° (20 mm.), and consisted of pure ethylic 

01820 gave 00669 AgBr. Br = 27-23. 
CuHjjBrO^ requires Br = 27' 

27*11 percent. 

Hydrogen Ethylic Salt of a-BroTno-PfidimcthylgluUbrie Acid, 
COOH- CHBr-C(CH8)2- CHg- OOOC2H5. 

This is precipitated in considerable quantity when hydrochloric 
acid is added to the sodium carbonate washings obtained as ex- 
plained in the last paragraph. It is a heavy, colourless oil which 
distils without decomposition at 240"^ (35 mm.). 

01932 gave 0*0771 AgBr. Br « 29*80. 

CgH^sBrO^ requires Br = 29-97 percent. 

The formation of this acid ethylic salt in the proportion of about 
20 per cent, of the total product of bromination is not due to the 
presence of water in the alcohol used, as is shown by the fact that 
exactly the same amount was formed in an experiment in which extra 
precautions were taken to eliminate, as far as possible, every trace of 
water. It seems to us probable that, from the examination of a 
number of similar cases, some acid bromides have the power of decom- 
posing alcohol in such a way as to form tha free acid and ethylic 
bromide, thus : R-COBr + HO-CjjHg = R-COOH + CgH^Br. 

It is also possible that the position of the bromine atom may 
scoount for the difficulty with which this hydrogen ethylic salt is 
further etherified, as an experiment which we made with the object of 
converting the hydrogen ethylic salt into the neutral ethereal salt, by 
means of alcohol and hydrogen chloride, showed that very little 
etherification had taken place. 

Methylie a-broma-fipdimetkylglutarate, 

8 obtained when the hromo-acid bromide of dimethylglutaric acid, 

„.gitized by Google 


prepared as explained above^ is poured into well cooled methylio 
alqphol. It is a mobile liquid boiling at 172° (20 mm.). 

Lactone of a-ffydroxy-PP-dimethylglutaric Acid, 


This interesting substance, which is isomeric with the caronic acids, 
was prepared as follows. 

The hydrogen ethylic salt of bromodimethjlglutaric acid (20 grams) 
was dissolved in dilute sodium carbonate and boiled in a reflux 
apparatus on a sand bath, care being taken that the solution always 
had a distinctly alkaline reaction. As soon as a small quantity of 
the liquid gave no precipitate with hydrochloric acid, the whole was 
acidified, saturated with ammonium sulphate, and repeatedly extracted 
with ether ; the dried ethereal solution, when evaporated, gave a hard, 
crystalline mass which, on being left in contact with porous porcelain 
for some days, became quite colourless and melted at 80 — 100^. In 
order to purify this crude product, and especially with the object of 
determining whether it contained any trana-caxonic acid (p. 59), the 
lactone was dissolved in a slight excess of ammonia, evaporated on 
the water bath, and the well dried ammonium salt warmed with 
absolute alcohol, when the whole dissolved readily, showing that no 
trana-oaxomo acid was present. Ether was then added to the alcoholic 
solution until a slight turbidity was produced, and the fine, transparent 
prisms of the pure ammonium salt deposited on standing were 
coUectr i, washed, and converted into the acid, which was further 
purified by repeated recrystallisation first from benzene and then from 
water. Analysis. 

01643 gave 03225 CO, and 00932 HjO. = 5353 ; H - 6-30. 
CyHi^jO^ requires C = 53-16 ; H = 6-33 per cent. 

This lactone melts at 112°, and when heated in a small retort 
distils without change. It is readily soluble in ether, acetone, ethylic 
acetate, and hot benzene, moderately in chloroform, and almost 
insoluble in light petroleum. 

Action qf Alcoholic PoUuh on Ethylic a-Bromo-PP-dimethylgluUiraU. 
FomuUion qf cis- and trans-(7aronio Acids, 



In this interesting experiment, the brom-ethereal salt (15 grams) 
was digested in alcoholic solution with caustic potash (15 grams) for 

Digitized by VjOOQIC 


10 hours, and the product, after being freed from alcohol bj cTapora- 
tion on tbe water bath with the addition of water, was diflsolyed in 
^ater, acidified, and extracted sereral times with ether. On distilling 
off the ether, a syrupy mass was obtained, which became partially 
solid on standing ; this was dissolved in a little water and saturated 
with hydrogen chloride, when the crystalline solid, which slowly 
separated after being collected and dried on a porous plate, melted 
indefinitely between 170° and 200°. This crude substance was dis- 
solved in ammonia, evaporated to dryness, and the residual solid 
ammonium salt ground up with cold absolute alcohol; a small 
quantity passed into solution, but most of it remained undissolved.* 
The insoluble salt was collected, washed with absolute alcohol, 
dissolved in a little water, acidified and extracted with ether ; the 
ethereal solution, on evaporating, deposited a solid acid which even 
before recrystallising melted at 210—212°, and after recrystallising 
from water at 213°.' This substance is trans-ooronte octi. The 
alcoholic filtrate from the insoluble ammonium salt of ^a9M-caronic 
add, on being mixed with ether and allowed to stand, deposited a 
small quantity of crystalline solid ; this, after collecting, acidifying, and 
extracting with ether, yielded an acid melting at 176°, which, doubt- 
less, consisted of cif-<»^ronic acid (m. p. 176°), but the quantity was 
too small for analysis. In this hydrolysis, therefore, both the 
caronic adds appear to be formed, but the ^an««modification in far 
larger quantity than the et9-modification. 

On evaporating the solution after the precipitation of the cis- and 
trwu-caxomc acids to dryness with ammonia, and dissolving the 
residne in alcohol, an oily ammonium salt was precipitated on the 
addition of much ether, which, on long standing, became solid ; from 
this, on acidifying, a considerable quantity of the lactone of hydroxy- 
dimethylglutaric acid was obtained, melting at 112°. 

Aeiwn of Pciaah en Oe Hydrogen Bthylie Salt qf a-Bromo-fifi-dimethyl- 

gkUarie Acid, 

This hydrolysis, which gives by far the best yield of tron^-caronic 
acid, was conducted as follows. Equal weights of the brom-ethereal 
salt and caustic potash were heated together in alcoholic solution in a 
reflux apparatus for 3 hours, and after evaporating the alcohol, 
acidifying, and extracting with ether, exactly as explained in the last 
experiment, a hard, solid, crystalline cake was obtained ; this, on 
being crystallised once from water, melted at 213°, and consisted of 
pure <r0tw^»ronic acid. On treating this acid and its mother liquors 

* This nit is the acid ammoniam salt of traTU-cawnic acid, and has the formula 
C»H],K04 (Baeyer, Ser„ 1896, 29, 2800). 

Digitized by VjOOQIC 


with ammonia and alcohol, as explained above, we were not able to 
extract even traces of the cif-modifioation or of the lactone of 
hjdroxydimethylglutaric acid. 

It is remarkable that hydrogen ethylic bromodimethylglutarate 
should behave so differently from the normal ethy lie salt on treatment 
with potash under the same conditions, and that in the latter case, 
besides ^anf-caronic acid and traces of the ct9-modification, such con- 
siderable quantities of the lactone of hydroxydimethylglutaric acid 
should be produced. 

The formation of ^roTW-caronic acid from the hydrogen ethylic salt 
seems to us to prove that this salt has the formula 

COOH- CHBr-C(CH3)j« OHg- COOC3H5. 
given to it on p. 50, and that elimination of hydrogen bromide takes 
place before hydrolysis. If this ethereal salt had the alternative 
formula, C00C2Hj-CHBr-C(C 0^3)2' CH2- COOH, the elimination of 
hydrogen bromide would be expected to take place between the 
bromine atom and the hydrogen atom of the carboxyl group, and the 
lactone of hydroxydimethylglutaric acid would be formed ; this, 
however, is not the case. 

Action of Dietkylanilvne on Ethylic Bromodimethylglutarate, — As it 
has frequently been found that diethylaniline is a very valuable 
reagent for removing hydrogen bromide from organic substances, 
it was thought that interesting results might be obtained if its 
behaviour were investigated in the present instance. Accordingly, 
50 grams of ethylic bromodimethylglutarate was boiled in a reflux 
apparatus with 75 grams of pure diethylaniline for 2 hours, and after 
cooling, the nearly solid product was treated with dilute hydrochloric 
acid, and the oil which separated extracted with ether. The ethereal 
solution was dried, evaporated, and the residual oil fractionated a 
great many times, first under reduced and then under the ordinary 
pressure. It was thus separated into two fractions which boiled at 
241° and at about 265—275° " 

The oil boiling at 241°, on analysis, gave the following numbers. 

0-1345 gave 0*3032 COj and 0-1087 H^O. C = 61-48 ; H = 8-98. 
CijFjgO^ requires = 61-68; H»8-41 per cent. 

Since this oil, on hydrolysis, yielded (ran«-caronic acid, it is evidently 
the ethereal salt of this acid. 

The fraction 265 — 275°, which was not analysed, gave, on hydrolysis, 
the lactone of hydroxydimethylglutaric acid, and is evidently the 
ethereal salt of this lactonic acid. These two substances were ob- 
tained in about equal quantities. 

Digitized by VjOOQIC 


tnnBrCaronic acid, COOH-(^ — (J3-C00H. 

H H 

The synthetical acid has properties identical with those described 
by Baeyer and Yiiliger {Ber., 1896, 29, 2800) as characteristic for 
the add from carone. It is sparingly soluble in cold water, but 
readily in hot water, and separates from its hot solution in prisms 
which melt at 213^. It is very sparingly soluble in ether, ben^ne, 
and cold water, and almost insoluble in chloroform and light petroleum. 

01504 gave 0-2913 COg and 00888 HgO. = 62-82 ; H =- 666. 
CyHio04 requires C » 6316 ; H « 6'33 per cent. 

The gilver saU is precipitated as a white, crystalline powder when 
silver nitrate is added to a neutral solution of the ammonium salt. 

0-2660gave0-2177 OO2,00547H2Oand01638 Ag. C = 22-33; H = 229 ; 
Ag- 67-82. 
CyHgO^Aga requires C=* 22-57 ; H=«2-15 ; Ag = 58-06 per cent. 

<raii»>Caronic acid does not give an anhydride when digested with 
acetic anhydride, but when heated with acetic anhydride at 220° it 
yields the anhydride of ei»«aronic acid (p. 61). That it is a saturated 
add is shown by the fact that its solution in sodium carbonate does 
not reduce permanganate. 

Conversion of trans-(7aronic Acid into Terebic Acid, 



It was stated in the introduction to this paper that one of the most 
remarkable reactions of trans- and cis-caronic acid discovered by 
Baeyer, was the transformation of these acids into terebic add by the 
action of hydrobromic acid at 100% and it was consequently of im- 
portance to show that the synthetical acids behaved in the same 
manner under the same conditions. 

About 0'6 gram of pure synthetical irans-cajconio add was heated 

with about 5 cc. of concentrated hydrobromic acid (saturated at 0°) 

for 5 hours at 100^; the hydrobromic acid was then removed by 

evaporation^ on the water-bath, and the residue recrystallised^^ from 

water ; the colourless, cubic crystals melting at 174° thus obtained gave 

the following numbers on analysis. 

0-U22 gave 02757 COj and 0-0802 HgO. C = 5287 ; H = 6-26. 
C-jHj^O^ requires C « 6316 ; H = 6-33 per cent. 

„.gitized by Google 


Terebic acid has the same empirical formula and the same melting 
point as ct9-caronic acid, but in other respects these acids possess very 
different properties, and there can be no doubt that the acid obtained 
in the above experiment was terebic add, and not unchanged cU- 
caronic acid, for the following reasons (compare Baeyer, Bw,, 1896, 
29, 2799). This acid yields an ammonium salt which differs from the 
ammonium salt of <»!8^oaronic acid in that, besides having a different 
crystalline form, its solution in alcohol is not precipitated by ether. 
It gives, with silver oxide, a silver salt which is readily 'soluble in 
water and crystallises in needles, and on boiling with^ baryta water 
it yields the characteristic crystalline barium salt of diaterebic acid. 
Finally, a small quantity heated in a test tube gave the odour of 
pyroterebio acid, 

?^;?H.^H -,.ch:c(ch3),+co, 

and the residue, dissolved in soda, instantly reduced permanganate, a 
behaviour not shown by ci8-caronic]acid, which, under these conditions, 
is simply converted into its anhydrida 

Conversion qf trMiB'Caronio Acid into ciB-Caronic Acid. 

This conversion, which had not previously been observed, may be 
readily accomplished in the following way. ^an«-Caronic acid is mixed 
with three times its weight of freshly distilled acetic anhydride, and the 
mixture heated in a sealed tube for 6 hours at 220^. The dark brown 
'product is then freed from the excess of acetic anhydride by distillation, 
the residue dissolved in boiling water, digested with animal charcoal, 
filtered, and evaporated to a small bulk ; on cooling, large, glistening, 
colourless crystals separate, which melt at 1 76^, and consist of pure 
cig-caronic acid. 

0-1602 gave 0-3108 CO, and 00900 HjO. = 52*91 ; H = 6-24. 
CyHjoO^ requires = 63-16 ; H = 6-33 per cent. 

cis-Coronio cusid^ HC OH , is very sparingly soluble in 

cold water^ but readily in hot water, and crystallises from water 
most beautifully in brilliant, glistening plates with bevelled edges. 
It is sparingly soluble in dry ether and in light petroleum, and practi- 
cally insoluble in chloroform ; it dissolves readily in sodium carbonate, 
and this solution does not decolorise permanganate. When heated 
above its melting point, ot9-caronic acid is rapidly converted into its 

Digitized by VjOOQIC 


anhydride. The ammonium salt of cM-caronic add is readily obtained 
by dissolving the acid in excess of aqueous ammonia and evaporating 
the solution on a water bath. The crystalline residue differs most 
sharply from the ammonium salt of the tram^&cLd in being readily 
soluble in absolute alcohol ; from its alcoholic solution, it is precipitated 
by ether in the form of slender needles. When cw-caronic acid is 
heated with hydrobromic acid, under the conditions given in detail in 
the CQrresponding experiment with the ^raTi^acid (p. 59), it is con- 
verted into ierebie add melting at 174^. 

It will be seen from this short description of the properties of ctB- 
caronic acid that the synthetical acid is identical with the acid obtained 
by Baeyer (loe. cU.) from carone. 


Anhydride of cis-Caronic add, HC — CH. — ^This anhydride is 

dx)— O— CO 

foimed either when ci^-caronic acid is distilled, or when trcms-caxonic 
acid is heated at 220^ with acetic anhydride, but it is best prepared 
by boiling ct»-caronic acid with acetyl chloride until hydrochloric acid 
ceases to be evolved, evaporating, and crystallising the residue from 
dry ether, wben lustrous plates are obtained which melt at 56^. 

There caim be no doubt that this is the anhydride of the cM-acid, 
because it is, as Baeyer found {loe. cit,, p. 2799), quantitatively con- 
verted into the acid on boiling with water. 

The authors wish to state that this research was carried out with the 
aid of a grant from the Royal Society Eesearch Fund, and that they 
are indebted to Mr. E. H. Lees for making most of the analyses given 
in this and the succeeding communication. 

Chmis CoLLsoE, 

VIL — Synthesis of afiB'Trimethylglutaric Add. 
COOH- CH(OH3)-C(CHj^8- CHg- COOH. 

By W. H. PsBKiK, jun., and Jooelyn P. Thobpb. 

Or the four theoretically possible trimethylglutario acids, two only, 
up to the present, have been prepared synthetically, namely, the aooi- 
add, COOH-C(CH3)j,-CHj-CH(0H8)-0OOH, by Auwers and Victor 
Meyer {Ber., 1890, ^, 293) from ethylio a-bromisobutyrate and '' mole- 
cular " silver, and the aa)8-acid, COOH- C(0H8)2- CH(0H8)-0Hj- COOH, 

Digitized by VjOOQIC 

62 pI:r^in and TfiOBPE: synthesis of 

which was prepared by us (Trans., 1897, 71, 1187), by the reduction 
of the corresponding trimethylglutaconic acid, 

COOH- C(CH3)2- C(CH3):CH- COOH. 
The remaining two acids are 

app COOH* OH(CH3)-C(CH8)2- CH,- COOH, and 

apay^ COOH- CH(CH3)-CH(CH3)-CH(CH3)-COOH. 

Of these, the first has a special interest, for the following reason. 

By the oxidation of camphoric acid with permanganate at the ordi- 
nary temperature, Balbiano {Ber., 1894, 27, 2133 ; 1897, 30, 1908), 
obtained an acid of the formula C^H^j^s* ^ which he assigned the 

COOH- C(CH3)- C(CH3)3- CH- COOH. 

I O ! 

This acid, on reduction, yielded a lactonic acid of the probable 

CH(CH3)-C(CH3)8- CH- COOH, 

and this, by further reduction with hydriodio acid, was converted into 
an acid, CgH^^O^, which Balbiano considered to be a^;3-trimetbyl- 
glutaric acid, because, on oxidation, it yielded oa-dimethylsuccinic 

Assuming that this acid is aj9)3-trimethylglutaric acid, its formation 
from camphoric acid is important, as affording evidence that the latter 
contains the group C-C(CH3).C(CH3)2-C'C, a conclusion which, to- 
gether with results obtained from the great amount of work which has 
been done on camphoronic acid, throws much light on the constitution 
of camphoric acid (compare Trans., 1898, 73, 797). 

It thus became an important matter to be certain that the 
formula Balbiano assigned to his acid is correct, and for this reason 
we have, ever since the publication of his paper, been endeavouring to 
synthesise ay3)9-trimethylglutaric acid, and it is only quite lately that 
we have been able to accomplish this. 

In the first place, we found, as the result of a great many experi- 
ments, that it seems to be impossible to introduce a methyl group 
into ethylic dimethylpropanetricarboxylate, 

(COOC,H5)2CH- C(CH3)j- CHj- COOCgH^, 

at the point marked *, by the action of sodium ethoxide and methylic 
iodide. Arguing from analogy to other mono-substitution products of 
ethylic malonate, this experiment should have yielded an ethereal salt 
which would have given ay3)3-trimethylglutaric acid on hydrolysis 

„.gitized by Google 


and elimination o! carbon dioxide, but> although tried under very 
varied conditions, no trace of this acid was obtained. 

It was then thought possible that the ethylic trimethylpropanetri- 
carboxylate, (COOC2H5)2C(CH3)-C(CH3)2- CHg- GOOC2H5, which should 
have resulted from the above experiment, might be formed by the 
condensation of ethylic dimethylacrylate with the sodium derivative 
of ethylic mebhylmalonate, bub the experiments made in this direction 
did not yield a trace of the desired compound, and many other syn- 
theses, which need not be mentioned here, also gave negative results. 
Ultimately, we discovered the following synthesis, which is so easily 
carried out that large quantities of aySyS-trimethylglutaric acid can 
now be obtained in a short space of time. 

The method consists in heating together ethylic dimethylacrylate 
and the sodium derivative of ethylic cyanaoetate in alcoholic solution 
until the condensation to the sodium derivative of etfiylic eyanodi- 
meihylgluiaraie, which probably has the formula, 

OOOC^Hj- 0(CN)Na- 0(OH3)2- CHg- COOOgHg, 
is complete (see p. 5^). 

The crude product is then directly treated with methylic iodide, 
when an almost quantitative yield of ethylic cyanotrimethylglutarate, 
COOO^Hj- C(CHj)(CN)- C(CH3)2- CH^- OOOOjHg, is obtained. This, on 
hydrolysis with methyl alcoholic potash or hydrochloric acid, yields the 
beautiful, crystalline imide of a)3)9-trimethylglutaric acid, 

which, when heated with hydrochloric acid, at 200°, is converted 
quantitatively into a/3)9-trimethylglutaric acid. 

A direct comparison of the acid thus synthesised with that obtained 
from camphoric acid was rendered possible by the kindness of Professor 
Balbiano, who sent us a sample of the acid which he was the first to 

Both acids melted at 87% and on treatment with acetyl chloride 
yielded the same anhydride melting at 82% and from this by the 
action of aniline an anilic acid was prepared which in both cases 
melted at 150 — 151% so that there can be no doubt as to the identity 
of the synthetical acid and Balbiano's acid. 

In experimenting with a)9)9-trimethylglutaric anhydride, it was 
found thaty not only does it crystallise from water unchanged, but 
that it actually crystallises with water of crystallisation, in prisms 
melting at 61% and without being converted into the acid. 

This is certainly a very unusual property of an anhydridei and the 
only somewhat similar case which we have been able to find is that of 
the ^lactone of dimethylmalic acid, (^^^W QH-OOOH ^j^.^j^^ 

Digitized by VjOOQIC 


as Baeyer and Yilliger {Ber,, 1897, 30, 1955) have shown, crystallifies 
with IHgO. 

We are at present engaged in investigating other condensations 
between ansaturated ethereal salts and a-cyano-ethereal salts, and also 
on the action of halogen compounds on the sodium derivative of 
ethylic cjanodimethylglutarate, and we are especially interested in 
the ethereal salt which this sodium derivative yields on treatment 
with ethylic bromacetate, because, if Baeyer's formula for isocam- 
phoronic acid be correct, this should lead to a synthesis of this very 
important acid. 

Ethylic a-GyanoHi^'Pp'trimeihylgluta/raUy 
COOCgH^- CH(CN)-C(CH8)2' CH(CH3)-COOC2H5. 

In order to prepare this substance, ethylic dimethylacrylate is 
digested in alcoholic solution with the sodium derivative of ethylic 
cyanacetate, as explained on page 52, and after heating for 15 hours 
excess of methylic iodide is added and the boiling continued until the 
liquid has a neutral reaction ; water is then added, and the oily product 
extracted with ether. The ethereal solution, after washing well with 
water, drying over calcium chloride, and evaporating, deposits a thick 
oil which, after twice fractionating, boils constantly at 181° (25 mm.), 
and consists of pure ethylic cyanotrimethylglutarate. 

0*22 gave 11*8 c.c. nitrogen at 23° and 750 mm. N»5'94. 
CjgHgiNO^ requires N = 5'50 per cent. 

The yield of this pure ethereal salt obtained in this condensation 
is about 68 per cent, of the theoretical, and it is remarkable that, in 
this case, only traces of an acid ethereal salt are formed, whereas, as 
explained on p. 52, if the condensation is worked up before the treat- 
ment with methylic iodide the principal product is the acid ethereal 

afi^'TrifMthylglutarimide, C(CH,),<gg| ^^«)'g g>NH. 

This substance is formed when ethylic cyanotrimethylglutarate is 
hydrolysed either with methyl alcoholic potash or with concentrated 
hydrochloric acid ; in the former case, the product often consists of 
the acid amide, a white, deliquescent solid which, however, can readily 
be converted into the imide by boiling for a short time with concen-- 
trated hydrochloric acid. a^/3-Trimethylglutarimide is sparingly 
soluble in cold, but readily in hot water, and crystallises in lon^ 
needles closely resembling )3)3-dimethylglutarimide (p. 53) in appear-- 
ance. It melts at 126°. 

Digitized by 



0-1163 gave 9-4 c.c. nitrogen at 22° and 766 mm. N = 9-20. 
CgHjjjNOg requires N =3 9-03 per cent. 

Silver Malt, C(CH3)2<25S5o'-^ interesting salt 

is obtained as a beautiful, crystalline precipitate when dilute ammonia 
is carefully added to a warm solution of the imide and silver nitrate. 

00874 gave 0-0363 Ag. Ag = 41 53. 

CgHjgOgNAg requires Ag» 41*22 per cent. 

appTrimethylgluUmc acid, COOH-CH(CH3)-0(CH8)2-CH2-COOH. 

This acid is best prepared from the imide just described by heating it 
with three times its weight of concentrated hydrochloric acid in a sealed 
tube for 5 hours at 200^ The product is evaporated to dryness on 
the water bath, and the residue extracted with ether ; after drying 
over calcium chloride, the ethereal solution deposits the acid in an 
almost pure condition on evaporation. For the analysis, the acid was 
crystallised from water, from which it separates in glistening plates 
melting at 88°. 

0-1316 gave 0-2675 CO^ and 00960 FgO. = 55-43 ; H«8-23. 
CgHi^O^ requires = 55-17 ; H = 8'05 per cent. 

aPP-Trtmethylglutaric Anhydride, 9f^(^58)'C(0H8),-^, _^j^.g^ 

which is the most characteristic derivative of aj9)9-trimethylglutaric 
acid, is readily prepared by digesting the pure acid for a short time 
with excess of acetyl chloride, and then evaporating to dryness on 
the water bath. The crystalline residue, which consists of the almost 
pure anhydride, crystallises from a mixture of ethylic acetate and light 
petroleum in prisms which melt at 82°. 

0-1446 gave 0-3248 00^ and 0-0991 HA = 61-34; H = 7-62. 
OgHigOs requires = 61-54 ; H = 7-67 per cent. 

This anhydride is quite insoluble in cold dilute sodium carbonate 
solution, and is only very slowly converted into the acid even on 

If this anhydride \a boiled with a large quantity of water, and the 
clear solution, rapidly poured off from the oily drops, is allowed to 
cool, a smaU quantity of the anhydride crystallises out in needles (J) ; 
the same substance is produced on melting the anhydride under hot 
water and stirring vigorously until it becomes solid (II). The pro- 
duct in both cases melts at 61°, and consists of the anhydride with 
half a molecule of water of crystallisation. The following are the 
results of the analyses of specimens of I and II. 

VOU MULV. Digitized by QoOgle 


I. 0'2322 gram, after drying on a porous plate for 3 hours, lost 

0134 gram at 50°. HgO = 5 -8 per cent. 

II. 0-4712, heated at 50° until constant, lost 00242. H20 = 615. 

CgHjgOg-t-JHgO requires H20 = 5'4 per cent. 

The water of crystallisation is given off rapidly on gently warming, 
and slowly over sulphuric acid in a vacuum desiccator ; the residue 
then melts at 82°, the melting point of the dry anhydride. 

afiP-TrimethylglttUbranUie Add, 

CeHg-NH- C0-CH(CH8)- C(CH3)2- CHj- C00H(1). 
On the addition of aniline to a solution of the anhydride in benzene, 
this compound separates after some time as a white, crystalline 
precipitate. After being purified by recrystallieation from dilute 
methylic alcohol, from which it separates in long needles, it melted 
at 150 — 151°, and, on analysis, gave the following result. 

0-2515 gave 12*2 c.c. nitrogen at 21° and 752 mm. N=5'47. 
Ci^HjgNOj requires N =5 5-60 per cent. 

Our thanks are due to Messrs. F. Howies and F. H. Lees, for much 
valuable help in connection with this and the preceding research, and 
we wish also to state that the very considerable expense which these 
experiments have entailed has been largely met by grants from the 
Royal Society Research Fund. 

OwBNs College, 

VIII. — Occurrence of Orthohydroxyacetophenone in the 
Volatile Oil of Chione glabra. 

By Wyndham R. Dunstan, F.R.S., and T. A. Hbmry, Salters* Research 
Fellow in the Laboratories of the Imperial Institute. 

Some years ago, one of us (Dunstan, Proc. Roy, Soc, 46, 211) showed 
that the strong foBcal odour of the wood of Celtis retictUosa is due to 
the presence of skatole, the substance to which the odour of human 
excrement is due. Through the interest of Mr. Thiselton Dyer, the 
Director of Kew Gardens, other plants having strong odours were 
then examined, but no definite information as to their constituents 
could be obtained, owing to the small quantity of material available 
for examination. Amongst these plants was a small specimen of the 
wood of Chione glahra, which had been sent to Kew by Mr. J. H. 
Hart, F.L.S., Superintendent of the Royal Botanic Gardens, Trinidad, 
with the suggestion that an examination of its constituents should be 

„ .gitized by VjOOQ , _ 


undertaken, as the plant is reputed to possess valuable properties as a 
medicine and, in particular, is stated to be a powerful aphrodisiac. The 
bark and wood emitted a strong, unpleasant odour, chiefly aromatic, 
but partly foecal. This was proved to be due to an oil which was volatile 
with steam, and although its chief physical and chemical properties 
were ascertained, the constitution of its principal constituent could 
not be determined owing to insufficiency of material. A further supply 
of the wood was afterwards obtained from the collections of the Imperial 
Institute. This was part of a large log which had been sent previously 
for exhibition at the World's Fair at Chicago. Now, however, it has 
completely lost its characteristic odour, owing to the escape of the 
volatile oil, and it proved to be quite useless for our purpose. Through 
the kind offices of Mr. Hart, the Trinidad Government undertook the 
collection of a fresh quantity of the wood, which has made it possible 
for us to complete the inquiry by identifying the odorous constituent 
of the tree, and in fact, proving its identity with the substance 
prepared synthetically in the laboratory. 

The genus Chume of the Natural Order Ruhtc^cetB includes plants 
which are confined almost exclusively to the West Indies. The species 
Chione glabra, which is indigenous to Grenada, is a large flowering 
tree, known in the island as *'Yiolette," a name which probably 
has reference to the aromatic smell of the flower. Through the 
kindness of Mr. E. J. Millard, F.C.S., we have received from Grenada 
a dried specimen of the stem, leaves, and flowers of this tree. The 
somewhat aromatic, somewhat foecal, smell is associated with the 
bark and wood, especially with the former ; on exposure to air, it 
gradually disappears. The results of the chemical investigation 
described in the subsequent part of this paper show that the volatile 
oil, which exhibits in a concentrated form the remarkable odour of 
the wood, is composed of two substances, the one a yellow oil boiling 
at 160^ under 34 mm. pressure^ and solidifying at low temperatures 
to a crystalline mass ; the other, a colourless^ crystalline substance 
melting at 82°. The former, which is the chief constituent, we have 
proved to ' be orthohydraxyacetophenone, OH-C^H^'CO'CHg, and is 
identical with the compound prepared synthetically. The crystalline 
sabetance, which is present in very small quantity, has all the proper- 
ties of analkyl derivative of the oil ; the amount of substance obtained, 
however, was insufficient to enable us to complete the examination. 
The foecal odour of the fresh wood, which seems rather more pro- 
nounced than that of the constituents we have isolated from the oil, 
suggested the possibility that the plant might also contain a minute 
quantity of skatole or some other derivative of indole ; but although 
the volatile oil contains traces of somejnitrogenous substance, we have 
not been able to isolate any indole derivative from it. It is, however. 

Digitized!) Google 


interesting, and possibly significant from the biological standpoint, to 
observe that derivatives of orthohydrozyacetophenone may, by 
processes involving condensation and elimination of water, pass into 
compounds belonging to the indole group ; the production of indoxyl 
from orthacetylamidoacetopbenone has, indeed, been accomplished by 
von Baeyer and Bloem {Ber,^ 1884, 17, 963), and it is not difficult to 
conceive that the conversion may occur without difficulty and by 
shorter steps in the plant. The precursors of skatole and other indole 
derivatives in plants are at present unknown, but the occurrence of 
derivatives of acetophenone in the vegetable kingdom and particularly 
in Chione ghhra^ whose wood emits a foecal odour, indicates at least 
one possible direction in which search might be mada 

Besides those whose names we have already mentioned, we are 
greatly indebted to Miss L. E. Boole, F.I.C., who conducted much of 
the preliminary examination of the constituents of this plant. 

Extraction qf the Volatile Oil. 

The volatile oil was separated from the wood and bark by cutting 
these into fine shavings and distilling with steam ; the distillate 
obtained was then shaken with ether, and the ethereal solution dried 
over calcium chloride and distilled. After removal of the ether, there 
remained a dark-coloured oil which distilled with some decomposition 
under the ordinary pressure ; it was, therefore, distilled under reduced 
pressure, when a fraction boiling from 160° to 165°, under a pressure of 
34 mm., was obtained. With the small quantity of material available, 
it was not possible to obtain a fraction of more definite boiling 

Combustions of the liquid gave the following results. 

0-0737 gave 01926 OOg and 00352 HjO. = 71-2; H = 5-29. 
00607 „ 0-1586 CO2 „ 0320 H^O. C = 71-21 ; H = 5-76. 
CgHgOg requires C = 7057 ; H = 5*88 per cent. 

The relative density of the oil is c^ = 0-850 1574°. 

It is slightly soluble in water, and has a peculiar and somewhat 
unpleasant odour, chiefly aromatic, but partly foecal. An aqueous eola- 
tion of ferric chloride produces a deep purple coloration, and bromine 
water a faintly yellow, crystalline precipitate. The oil dissolves in 
alkaline solutions, and such solutions, on evaporation, leave a crystal- 
line salt or metallic derivative ; the potassium and sodium salts crys- 
tallise in yellow plates which quickly decompose on exposure to air ; 
acids regenerate the oil from them. 

Digitized by VjOOQIC 


Action of various Reagents on the Oil. 

Action of Acetic Anhydride.— The liquid was mixed with acetic an- 
hydride and warmed. The viscous oil precipitated on the addition of 
water to the product was removed by shaking with ether, the 
ethereal solution dried, and the ether removed by distillation. The 
crystalline residue of acetyl derivative obtained on standing was then 
recrystallised from methylic alcohol until the melting point was con- 
stant at 88^. A combustion gave the following result. 

0-1331 gave 03254 COg and 00776 HgO. C = 6666 ; H = 64. 
CgHyO-CgHjOa requires = 67-4 ; H = 5-61 per cent. 

The substance is, therefore, a monacetyl derivative of the original 
liquid, and the existence of one hydroxyl group is thus proved. 

Action of HydroQcylamins and Flienyl hydrazine, — When an aqueous 
solution of hydroxylamine is added to the oil suspended in water and 
the mixture gently warmed, a red, resinous substance separates ; this 
dissolves in boiling water, and the solution deposits colourless needles 
as it cools. These, after recrystallisation, melt at 112^, and have all 
the properties of an oxime of the original substance. 

0-1841 gave 0-4242 COg and 0-1016 H^O. = 628; H = 612. 
OH-CgH^-NOH requires C = 63-4 ; H = 5-96 per cent. 

A phenylhydrazone was also obtained by the action of a 10 per cent. 
solution of phenylhydrazine on the liquid suspended in water ; after 
24 hours, this mixture deposited a greenish-yellow oil which soon 
became crystalline. On recrystallisation, the hydrazone melted at 108^. 

The production of an oxime and hydrazone indicate the presence of 
one ketonic group in the original substance. 

Action qf Bromine. — When bromine water is added to the oil 
suspended in water, a colourless precipitate forms immediately, and on 
standing becomes crystalline ; after recrystallisation from hot alcohol, 
it melts somewhat indefinitely at 108°, or, after prolonged heating at a 
lower temperature, at 95°. We are] unable, at present, to represent 
its composition by any formula derivable from that of the parent sub- 
stance. Oombustions and two determinations of bromine by Oarius' 
method gave the following results. 

C. 32*6 percent. H. 2*52 per cent. Br. 4564 per cent. 
32-3 „ 2-04 „ 46-2 

32-96 „ 209 „ 

32-89 „ 2-26 „ 

It will be seen that these values for carbon, hydrogen, and bromine 
do not agree with those required for a mono- or dibromo-derivative. 

„.gitized by Google 


CgHyOaBr requires = 44-6 ; H = 3-2 ; Br = 37'2. 
CgHgOgBrg „ C = 32-6 ; H = 204 ; Br = 544. 

When the bromine derivative is reduced with tin and hydrochloric 
acid, it yields a small quantity of a phenolic substance giving a bromine 
derivative melting at 86^. It is possible that the solid bromo-deriva- 
tive is a mixture of one or more simple bromine derivatives, but 
this is not very probable, since its composition seems to be constant 
when prepared under various conditions, and its melting point is very 

Action of Fused Potash. — When the oil is boiled for some time (in a 
vessel fitted with an upright condenser) with 10 per cent, potash 
solution, it merely passes into a metallic derivative, and can be re- 
covered by acidifying the solution; and even when heated with alkaline 
solutions in sealed tubes at 125^, it is not decomposed. When fused 
with excess of potash, the potassium salt of the oil floats for a time on 
the excess of potash, but is subsequently completely decomposed, 
forming a dark brown mass. When this is dissolved in water, 
acidified, and the mixture distilled, a solution containing phenol is 
obtained ; this was identified by means of its tribromo-derivative. 
When the acid liquid left after distillation was extracted with ether, it 
furnished a crystalline substance which, after recrystallisation from 
hot water, melted at 151°, and gave all the reactions of salicylic add 
m. p. 155°). 

Action of Nitric Add, — ^The residue from the Carius* estimations of 
bromine in the bromo-derivative was examined, and found to contain 
oxalic acid and a yellow, crystalline substance j the quantity of the 
latter, however, was too small to admit of satisfactory identification, 
but since it dyed silk yellow and did not dye cotton, and gave the 
chloropicrin reaction, there can be little doubt that it was picric acid. 

These observations as to the chemical behaviour of the volatile 
oil of Chione glabra prove that its chief constituent is orthohydroxy- 
acetophenone [OH : C0-CH8= 1 : 2]. 

The facts may be conveniently summarised here. 

1. The parent substance has a composition represented by the 
formula CgHgOg. 

2. One of the hydrogen atoms can be replaced by the acetyl 
group, indicating the presence in the parent substance of one hydrox jl 

3. The oil forms a monoxime, having a composition represented by 
the formula OgHgO^N. 

4. Fusion of the oil with potash furnished salicylic acid and phenol, 
the latter as the result of secondary action. 

Digitized by VjOOQIC 


The formation of all these substances is readily explained on the 
assumption that the chief constituent of the oil is orthohydroxy- 
acetophenone, OH- CgH^-CO'CHg. This compound has been described 
by Tahara {Ber., 1892, 25, 1306) and by Eeuerstein and Kostanecki 
(Ber., 1898, 31, 710—719). The latter chemists described ortho- 
hydroxyacetophenone (obtained by the action of alcoholic soda on 
phenacylidenflavene) as a yellow oil of peculiar odour, distilling at 
218% giving a purple red colour with ferric chloride solution, and 
forming a yellow, crystalline, sodium salt. Tahara prepared the com- 
pound by the decomposition of orthomethoxybenzoylacetic acid. The 
acetyl derivative melted at 89% and the phenylhydrazone at 107^. 

We have prepared oxthohydroxyacetophenone from nitrocinnamic 
acid, by the following series of reactions. Orthonitrocinnamic acid, 
NOj'CgH^'CHICH'COOH, was converted successively into dibromonitro- 
phenylpropionic acid, NOj-CgH^-CHBr-OHBr-COOH ; orthonitro- 
propiolic acid, NOg'CgH^'CSC'COOH ; orthonitrophenylacetylene, 
NOg-CgH^'CiCH; orthamidophenylacetylene ; NHg-OgH^-CSCH; ortho- 
amidoacetophenone, NHg'C^H^'OO'CHg ; and, finally, by diazotising 
into orthohydroxyacetophenone, OH'CgH^'CO'CHg. 

These compounds, with the exception of the last, have been prepared 
by Baeyer and Bloem {Ber., 1880, 13, 2259), and we have made use 
of the methods of preparation and purification described by these 
chemists. The orthohydroxyacetophenone thus prepared had all the 
characteristics of the oil distilled from Chione ghhra, and furnished 
derivatives having the same physical constants; thus the bromine 
derivative and the oxime melted at 108° and 112° respectively. 

Orthomethox7/<i€etophenone. — ^The quantity of colourless, crystalline 
substance left as a residue in the distillation of the crude chione oil 
was too small to admit of an extended examination, but since it was 
decomposed by boiling with hydriodic acid, giving a volatile iodide, 
with the production of an oil having the properties of orthohydroxy- 
acetophenone, it is probably an ether, possibly the methyl ether of 
this substance. The minute amount of this constituent present in the 
volatile oil has made it impossible for us more certainly to identify it.* 

SciBHTiyic Depabtmekt, 
Impebial Institute, 
lokdon, s.w. 

* Since the work described in the present paper was completed, the following 
account of the Tolatile oil of Chtane glabra has appeared (Paul and Cownley, Fharm. 
Joum. , [i v], 7» 51). The material examined was obtained from the Windward Islands. 
"From the aromatic odour of the bark, it was conjectured that it might contain a 
Tolatile oil, and on examination with that object we succeeded in isolating a Yolatile 
oQ amounting to about 1 '5 per cent, by weight. 

" This yolatile oil is of a pale yellow colour, and has a specific gravity higher than 

„.gitized by Google 


IX. — Occurrence of Hyoscyamine in the Hyoscyarrvus 
muticus of India. 

By Wyndham R. Dunstan, F.R.S., and Harold Brown, Assistant 
Chemist in the Laboratories of the Imperial Institute. 

Hyo$cyamu8 muticus is a species of henbane which occurs in certain 
districts of India, and has long been used in Indian medical practice, 
as a particularly virulent drug. The nature of its alkaloid, however, 
has never been determined. 

The following account of the plant is abbreviated from that given 
in Hooker's Flora Indtca. 

ff, muticus is found in the West Punjab, in Scinde, and in Cabul, 
westward to Egypt. 

Oauline leaves, petioled, ovate or oblong, entire or toothed, lower 
flowers pedicelled, calyx striate, pubescent, teeth short, triangular, 
not acute in fruit, corolla 1 to 1^ inches, lurid yellow or nearly 
white. Stem, 1 to 3 feet. Leaves, 4 to 7 inches, pubescent, or some- 
what woolly, petiole ^ to 3 inches. Lower pedicles in fruit, j^ to 1 inch. 
Calyx, § inch, in fruit, 1 by ^ inch, funnel-shaped, ribbed, somewhat 
reticulate, villous, or ultimately glabrous ; teeth short, triangular, not 
spreading. Capsule, ^ inch in diameter. Seed, ^^ inch in diameter. 

The medicinal effects of this drug appear to resemble those of 
ordinary henbane ; when administered in comparatively large doses, 
it acts as a powerful excitant, leading to what have been described as 
paroxysms of mania, whence the synonym Hyoacycmvus insanus. There 
is some reason to believe that it has been employed in the preparation 
of certain extremely potent kinds of the Persian " Benj " and in some 
varieties of the Indian *' Bhang," which, however, is usually prepared 
from Indian hemp. Since there is at present a considerable demand for 
atropaceous drugs and their alkaloids, it seemed desirable to examine 

that of water. It gave a mass of acicular crystals on being cooled to about - 20^. 
The oil gave no crystalline compound with sodic bisulphite, nor could evidence be 
obtained of methyl salicylate. It gave a semi-solid mass on treatment with a con- 
centrated solution of caustic soda. It is readily soluble in dilute caustic soda, and 
reprecipitated by acids apparently unaltered. It is slightly soluble in water, its 
aqueous solution gives a purple-blue coloration with ferric chloride. The volatile 
oil is evidently a phenol. 

** There are several volatile oils having somewhat similar properties to those above 
described, but the quantity of oil obtained was too small to admit of the further 
examination which its peculiarity appears to deserve." 

Digitized by VjOOQIC 


the alkaloid of this plant, which is fairly abundant in the districts of 
India in which it is found, and might possibly prove to be worth 

Extraction and Estimation of the Alkaloid, 

The sample, which weighed nearly two pounds, was collected in 
Scinde, at the instance of the Reporter on Economic Products to the 
Government of India, and consisted of both stems and leaves of the 
plant ; these were separated as far as possible, but there remained a 
mixture of fine stems and broken leaf. 

/ 300 grams leaf. 
Total weight, 900 grams... < 310 „ stem. 

( 290 „ mixture of ^leaf and stem. 

These three portions were extracted separately, in order to deter- 
mine the proportion of alkaloid in each. 

The material was air-dried at a low temperature, finely powdered, 
and then exhausted by percolation with cold alcohol, the percolate 
being evaporated under reduced pressure until nearly the whole of the 
alcohol had been removed. The semi-solid residue was then extracted 
with dilute hydrochloric acid (0*5 per cent.), gently warmed, and well 
shaken; after subsidence, the acid liquid was poured o£E and the 
treatment repeated until the alkaloid was completely removed. The 
acid solution, after filtration, was shaken with small quantities of 
chloroform in order to remove the chlorophyll, then made slightly 
alkaline by the addition of dilute ammonia, and the alkaloid extracted 
by successive shakings of the alkaline liquid with chloroform ; on dis- 
tilling off the chloroform under reduced pressure, a gummy, alkaloidal 
residue was obtained having a slight odour recalling that of pyridine. 

A small quantity of alkaloid remained in the alkaline liquid, and 
could not be removed by agitation with chloroform ; it was, however, 
subsequently isolated and examined. 

By percolation with cold alcohol, the following percentages of alkaloid 
were found. 

In the plant (stem and leaf), 680 grams yielded 0*7 gram of alka- 
loid, or 0*1 per cent (nearly pure). 

In the stem only, 310 grams yielded 0*46 gram of alkaloid (nearly 
pure), which is 0*15 per cent. 

In the broken leaf and fine stems, 250 grams yielded 0'14 gram of 
alkaloid (nearly pure), which equals 0*056 per cent. 

Examination of tlie Alkaloid, 

In order to identify the alkaloid or alkaloids present, the impure 
amorphous residues were converted into the aurichloride, which was 

„.gitized by Google 


examined by fractional crystallisation, as described in a previous paper 
on the alkaloids of Soopola Camiolica (Dunstan and Ohaston, Ph. /., 
[iii], 20, 461) ; this was obtained as a yellow, pulverulent precipitate, 
which was recrystallised from hot water acidified with hydrochloric 
acid. The alkaloid from the mixed stem and leaf yielded two large 
fractions of auric chloride melting at 154 — 156^ and, after recrystal- 
lising, at 167 — 168°, which is very near the melting point of hyos- 
cyamine aurichloride. From the mother liquors, containing the 
excess of auric chloride, two very small fractions of crystals were ob- 
tained melting at 149 — 160° and 140 — 146° respectively ; these were 
too small to be dealt with separately. 

The alkaloid from the stem yielded three main fractions of auri- 
chloride all melting between 164° and 156°. On recrystallising once, 
three fractions melting at 168°, 167'6° and 164 — 166° were obtained ; 
by concentrating the mother liquors from these, two small fractions 
melting at 160 — 165° and 146 — 160° were obtained, and, finally, a few 
crystals melting at 141 — 145°. The fractions of similar melting 
points were combined and recrystallised from dilute acid until a constant 
melting point was attained. In the case of the fractions of higher 
melting point, two or three recrystallisations sufficed, whereas 
the lower fractions required the operation to be repeated at least 
six times before a constant melting point was obtained ; by this 
means, practically the whole of the aurichloride originally taken 
was obtained in shining leaflets melting at 159*6°. There only 
remained two small fractions melting at 167 — 158° and the two 
lowest fractions, at 140 — 145°, which were too small to be separately 
dealt with, and were consequently retained and afterwards combined 
with a fraction of similar melting point. 

It is, therefore, proved that nearly the whole of the aurichloride is 
that of hyoscyamine, which melts at 160° (Ladenburg). 

The small quantity of alkaloid which remained in the alkaline 
liquid and was not removed by agitation with chloroform was isolated in 
the following manner (Dunstan and Kansom, Fh. J., [iii], 14, 623). The 
chloroform and alcohol present were removed by distillation under re- 
duced pressure, the liquid was acidified, and the whole of the alkaloid 
precipitated by adding a solution of iodine in potassium iodide ; the 
alkaloids! periodide was collected, washed, and then decomposed by a 
small quantity of a solution of sodium thiosulphate, the liquid 
made alkaline with dilute ammonia, and the alkaloid extracted with 
chloroform. The small quantity of alkaloid thus obtained was con- 
verted into the aurichloride, but this could not be recrystallised and was 
mixed with the lowest melting fractions (above referred to) for further 

Digitized by VjOOQIC 


Analt/818 qf the Aurichloride, 

The pure aurichloride meltiDg at 159 '5°, was analysed, the gold, 
chlorine, and alkaloid, being directly determined. A weighed 
quantity of the salt was dissolved in water, the gold precipitated as 
sulphide, ignited, and weighed as metal. After the hydrogen sulphide 
had been removed from the filtrate by passing a current of air, it 
was exactly neutralised with soda, and the chlorine determined by 
titration with centinormal silver nitrate. The filtrate from the silver 
chloride was concentrated in a vacuum over sulphuric acid, then 
made alkaline with dilute ammonia, and the alkaloid removed by 
agitation with chloroform, and the residue obtained on distilling off 
the chloroform was dried in a vacuum over sulphuric acid until its 
weight was constant. 

I. 0-0636 aurichloride gave 0*0201 Au, and required 422 c.c. 
N/100 AgNOg ( = 0-0003416 CI). Au = 31-60 ; Cl = 22'66. 

11. 0*0424 aurichloride gave 0*0134 Au, and required 27*8 c.c. 
N/100 AgNOy Au = 31*60 ; CI = 22*40. 

The t:wo solutions containing the alkaloid from both the above deter- 
minations were mixed, and the total amount of alkaloid estimated. 

0106 aurichloride gave 0*0491 alkaloid » 46*32 per cent. 

Gold. Chlorine. Alkaloid. 

Found 31-60, 31-60 2266,22*40 4632 

Calculated for CiyHgjNOj^HAuCI^ 31*34 2*2-54 45-95 

Another specimen of aurichloride yielded, on ignition, 31*25 per 
cent, of gold. 

From more of the pure aurichloride, the alkaloid was regenerated 
by the method above described, and after recrystallisation from a 
mixture of dry chloroform and petroleum, it was obtained in long, 
silky needles. These were dissolved in absolute alcohol, and the 
specific rotatory power determined. The solution was Isevorotatory. 

a=:3r. 1=1 dm. c = 2-04. « = 19°. 

Other observers have recorded numbers lying between 20^ and 21°. 

Hyoscyamus muticus as a Commercial Source of Pure Hyoacyamine, 

Having thus ascertained that the alkaloid in Hyoecyafmis muticus 
IB chiefly, if not entirely, hyoscyamine, experiments were made to ex- 
tract the crystalline alkaloid direct from the plant. For this purpose. 

Digitized by VjOOQIC 


about 1^ lbs. of the plant, both stem and leaf, was finely powdered 
and the alkaloid extracted by the method already described ; in this way, 
0*7 gram of impure alkaloid was obtained as a coloured, gummy 
residue, which could not be crystallised either from ether or from 
dilute alcohol. By dissolving in dry chloroform, however, and gradu- 
ally adding light petroleum, the colouring matter was precipitated, 
together with a little of the alkaloid, and a nearly colourless solution 
was obtained, from which, on adding more petroleum and allowing it 
to stand, groups of small needles were gradually deposited. 
Several fractions of the crystalline base were thus obtained, and were 
recrystallised in the same manner. Finally, pure hyoscyamine was 
obtained in colourless, silky needles melting at 105^. 

Absence qf other Mydriatic Alkaloids. — The small quantity of alka- 
loid precipitated along with the colouring matter by the first additions 
of petroleum to the solution in chloroform was also examined. The 
precipitate was extracted with very dilute hydrochloric acid, the 
solution filtered, and aurichloride added, when a yellow precipitate was 
produced which quickly aggregated to a resinous mass (fraction A) ; 
this was removed, and on adding more of the reagent to the filtrate a 
yellow^ pulverulent precipitate was obtained, which did not aggregate 
on standing (fraction B). These two fractions were recrystallised 
separately from dilute acid ; from fraction A, two crops of crystals were 
obtained melting at 154 — 155° and 156 — 158° respectively; and from 
fraction B, crystals were obtained melting at 157 — 158°. The mother 
liquors from all these were concentrated in a vacuum, and thus a few 
crystals melting at 146 — 149° were separated. 

All fractions of aurichloride insufiicient for separate recrystallisa- 
tion which had been obtained in the various experiments made were 
finally combined into two groups according to their melting points, 
155 — 158° and 140 — 145° respectively. These were recrystallised as 
before and virtually the whole of the gold salt was found to melt 
constantly at 169'5° A very small fraction melted at 158 — 159°, 
and another from 150—155°, but these were too minute for further re- 
crystallisation. The following aro the melting points of the auri- 
chlorides of the chief mydriatic alkaloids : atropine, 136 — 138°, hyos- 
cyamine, 159 — 160°; scopolamine or hyoscine, 198 — 199°. 

There is, therefore, no evidence of the existence in this sample of 
Hyoscyamus muticus of any mydriatic alkaloid than hyoscyamine. We 
believe that the plant will prove to be an important source of this 
alkaloid, since it can be isolated from it with far less difiiculty than 
from ordinary henbane (Hyoscyamus niger), which also contains the 
alkaloid hyoscine, and often atropine in addition. 

It should, however, be remembered that the nature of the alkaloids 
contained in atropaceous plants appears to vary with age, mode of culti- 

„.gitized by Google 



yation, and other circumstances. Thns, in belladonna, the proportion of 
hyoscyamine to atropine fluctuates widely with the age of the plant. 
It is, therefore, desirable that other specimens of Hyoacyamua muticua 
should be examined, in order to ascertain whether hyoscyamine is 
invariably the only alkaloidal constituent. 

It will be convenient to summarise here our present information as 
to the occurrence of hyoscyamine in various plants. 

HyoseyamuB niger and cdbua (hyoscyamine and scopolamine). 

Hyoaeyamua muticua (hyoscyamine). 

Atropa heUadonna (atropine, hyoscyamine, and a little scopolamine). 

DiUv^a sifrcMwnium „ „ ^ 

DuMsia myoporoides (hyoscyamine and a little scopolamine). 

Scopola camiolica (hyoscyamine). 

Scopola japonica (atropine, hyoscyamine, and scopolamine). 

Lactuea aativa and viroaa (hyoscyamine). 

The following percentages of total alkaloid have been recorded. 


Hyo9cyamu8 niger. 




belladonna, stramonium. 






0-21-0-41 0-15 

0-155— 173 


LeaTes 30-0 90 I^^^^Q.^gj 

ristyr. 0089— 0-069 
\2nd„ 0-045—0 068 


stem Xf^.i 


16— 0-37 

0-058— 0-1 


Stem . 











We are indebted to Mr. E. A. Andrews, Junior Assistant in these 
Laboratories, for the skilful help he has given in the conduct of this 
experimental work. 

Scientific Depabtment, 

Impbbial Institutb, Londoh. 

X, — Preparation of Hyponitrite from Nitrite through 

By ErwA&D Divebs, M.D., D.Sc, F.R.S., and Tahemasa Haga, 

D.Sc, F.C.S. 

SuLPHONATiON, followed by hydrolysis, readily converts an alkali 
nitrite into the unstable hydroxyamidosulphonate, which, in a solution 
saturated with potassium hydroxide, has all its hydrogen replaced by 

„.gitized by Google 


potassium, and is simultaneously resolved into hyponitrite and 

2HO-NH-S08Na + 4K0H = (KON)^ + 2KNaS08 + iRfi. 

This remarkable change was made known by us in 1889 (Trans., 
1889, 55, 760), along with the fact that it afforded a rich source of 
hyponitrite, since the quantity of the latter formed is equivalent 
to at least half the nitrite taken. But this, it was pointed out, de- 
pended on our method of preparing hydroximidosulphonate being 
employed, and as the description of that method was not published 
until five years later, the new source of hyponitrite could not be used 
advantageously by others. Even since then, the method has escaped the 
notice of seven separate workers on hyponitrites, namely, Thum, W. 
Wislicenus, Paal and Eietschmer, Tanatar, D. H. Jackson, Piloty, and 
Hantzsch and Kaufmann, to be at last, however, taken up by 
Kirschner (Zeit. anorg, Chem., 1898, 16, 424). Indeed, it may be said 
to have been rediscovered by Piloty, who in a paper on " an oxidation 
of hydroxylamine by benzenesulphonic chloride" {Ber,, 1896, 29, 
1559), describes the resolution of benzenesulphonic hydroxylamide 
by potassium hydroxide into hyponitrite and benzenesulphinate, with 
a yield also of half the calculated quantity of the former salt. 

As we have each of us very frequently employed the hydroxyamido- 
sulphonate method of preparing hyponitrite, and the details requisite 
in order to get high yields have not been published, it seems it will be 
of service to chemists to describe the method fully.* So far as our 
experience goes, we can be certain of getting a yield of 60 per cent, of 
the theoretical, and occasionally much more, although we have not 
succeeded in discovering the cause of this. In what follows, we assume 
that the convenient quantity of 2 decigram-molecules of sodium^ 
nitrite is taken, from which about 17 grams of silver hyponitrite may 
be obtained. 

In order to limit the quantity of potassium hydroxide required, 
which is very large in any case, no more water than is necessary must be 
used, and except that particular attention must be given to this point, 
the process begins exactly like that of making hydroxylamine sulphate 
from nitrite (Trans., 1896, 69, 1665). In a tared, wide-mouthed, 
round-bottomed flask of 200 — 250 c.c. capacity, 14'4 grams of 96 per 
cent, sodium nitrite, f together with sodium carbonate containing 10*6 
grams of anhydrous carbonate, are dissolved by heat in enough 
water to make the total weigh 83*5 grams. Some lead salt 

* Eirschner's method is an excellent form of onr process, but I prefer that 
described in this paper. — E. D. 

t But preferably, 18*8 grams of pure sodium nitrite, this being now very easy to 
prepare (this vol., p. 86). 

Digitized by VjOOQIC 


in the nitrite is deposited, bat goes into solution when the potassium 
hydroxide is added and gives no trouble. Sodium carbonate of any 
hydration may be used, but as, subsequently, more of this salt has to be 
added, and then should be approximately the monhydrate, it is con- 
venient to use this form throughout. Such a carbonate, almost pure, 
is generally found in the " dried " pure carbonate of commerce. 
Keeping the flask in active motion in an ice-and-brine bath, sulphur 
dioxide is passed in until a short time after the temporarily precipi- 
tated acid carbonate has redissolved, and just when a bit of lacmoid 
paper in the solution becomes fully red. About 0*1 c.c. strong 
sulphuric acid is then dropped in. If the temperature is kept below 
0°, the conversion of the nitrite into hydroximidosulphonate is com- 
plete, whereas if it rises much above 0^ some nitrilosulphonate would 
form, and interfere with the result. If higher temperatures have 
been avoided, the nitrite and carbonate taken in molecular propor- 
tions, and the sulphur dioxide not used in excess, the solution is ready 
to be hydrolysed ; but as the salts may not have been in exact propor- 
tion, and sulphur dioxide may also be present, it is best to blow a 
strong current of air through the solution at 0^, so as to expel any 
sulphur dioxide or nitric oxide that may be present. 

After this treatment, the solution is brought to about 30°, in order 
to start hydrolysis, and set aside for a day in a warm place with the 
flask corked. Complete hydrolysis to hydroxyamidosulphonate with- 
out further hydrolysis to hydroxylamine is thus secured,* and, conse- 
quently, just the calculated quantity of sodium carbonate (10'8 grams 
anhydrous) is found to be required, including that for the three or 
four drops of sulphuric acid added. The approximately monhydrated 
carbonate, added in fine powder, is briskly shaken with the solution, 
so as to hinder its caking, the last portions being dissolved by 
warming the flask. The solution of hydroxyamidosulphonate and 
sulphate, thus prepared, contains almost exactly half its weight of 
water, is therefore supersaturated, and is the strongest solution 
practically obtainable; it will be found to approach closely 113 grams 
in weight, so smoothly do the reactions proceed. The solution may 

* That the production of hydroxylamine is avoided if this detail is attended to, 
has been proved by shaking the solution, after it has been made alkaline, with 
sodium amalgam, which readily converts hydroxylamine into ammonia, but does not 
act on the hydroxyamidosulphonate. On testing in this way, no odour of ammonia 
can be renognised, and moist red litmus paper held in the bottle is barely affected. 
Eirschuer, using potassium hydroximidosulphonate, had to heat to boiling to effect 
hydrolysis, which is difficult then to complete without some of the hydroxyamido- 
sulphonate passing on into hydroxylamine. When the hydrolysis is incomplete, 
nitrite will be regenerated later on by the alkali ; and when it is earned too far, 
nitrite will also be formed during the oxidation of the hydroxylamine by the silver 
(or mercury) oxide.- 

Digitized by VjOOQIC 


be made weaker, and then concentrated, but the adjustment is 
troublesome, and the formation of hard cakes of sodium sulphate, 
which interferes with the proper working at the next stage of the 
operation, is difficult to avoid. 

The contents of the flask are now well drained into a basin, prefer- 
ably a hemispherical nickel basin, or, lacking that, a stout porcelain 
one, capable of holding about 500 c.c. Potassium hydroxide, free from 
chloride, assayed for real alkali and for water, and having not less 
than two-thirds and not more than 1 mol of water to 1 of the hydroxide, 
is required, for if it were anhydrous it would cause much heating and 
consequent decomposition of the salts; generally, the potassium 
hydroxide purified by alcohol and the more translucent varieties of 
stick potash contain about the right proportion of water, and 
dissolve in water without much rise of temperature. From 130 to 165 
grams of it, according to its degree of hydration, are quickly 
crushed in a warm mortar, thrown into the solution in the basin, 
and incorporated with it by means of a pestle. There is marked 
heating only just at first, which is better met by keeping the basin in 
water or resting it on snow or pounded ice for a very short time. On 
stirring in the potassium hydroxide, the solution sets to a stiff paste, 
if kept cold, quickly becoming thin again on further stirring, but full 
of an opaque, white precipitate of sulphate. If the basin has been 
cooled, hardly any gas escapes at first, but gentle effervescence and 
much frothing occur before long in any case. When the potassium 
hydroxide has been all ground up and dissolved, the basin is placed 
under close cover from atmospheric moisture and carbonic acid, and 
left in a warm place for 30 hours ; if kept for more than 50 hours, the 
quantity of hyponitrite sensibly, but slowly, diminishes. As much 
even as one-fourth of the hydroxyamidosulphonate may sometimes, in 
cold weather, still be present, and can be partly destroyed by keeping 
the basin at 55 — 60° for half an hour, although not with noticeable 
increase of the quantity of hyponitrite, but this heating, with the 
attendant risk of over-heating, is better omitted, on the whole. 
Besides undecomposed hydroxyamidosulphonate, the contents of the 
basin now consist of precipitated sulphate and sulphite, and solution 
of potassium hydroxide in slightly less than its weight of water 
(almost exactly KOH : 3H2O), together with the potassium hypo- 
nitrite. It is, apparently, only to secure this concentration of the 
potassium hydroxide, a practically saturated solution, that hardly less 
than 10 mols. of it to 1 of hydroxyamidosulphonate have to be used. 
More of it may be added without effect, good or bad, unless the solu- 
tion of the salts is weaker than is here recommended, for in that case 
additional potassium hydroxide must be used to bring the concentra^ 
tiontothe right point. 

Digitized by VjOOQIC 


Treatment with a silver salt is the only way of separating the 
hyponitrite from the other salts, and for this purpose the presence of 
the alkali is essential, together with large dilution when precipitating. 
The best way is to use the silver solution exceedingly dilute, because 
this checks the precipitation of silver oxide and sulphite until some 
time after all the hyponitrite has separated. Now the necessity 
for large dilution, and the advantage of still larger dilution, remove 
the only objection that can be raised to the use of silver sulphate 
instead of ^ver nitrate, and since it is generally important to feel 
assured that no trace of nitrate or nitrite can have been carried down 
with the hyponitrite, the sulphate should have the preference, although 
the nitrate can almost certainly be used with as good results. A cold 
saturated solution contains only 5 or 6 grams of the sulphate to the 
litre, and is most easily prepared by boiling excess of the salt with 
water and pouring the solution into an equal volume of cold water. 

Whichever salt is used, the contents of the basin are first washed 
into a very capacious precipitating vessel, and the highly dilute silver 
solation is poured in until it ceases to produce any more black 
precipitate. When this is at all abundant, as it sometimes is in 
cold weather, an hour's interval is given for subsidence of most of 
it, the still dark solution is decanted, and the precipitate washed by 
decantation before rejecting it. With or without this interruption, 
the addition of the silver solution is continued until the bright yellow 
hyponitrite suddenly appears, and so long after as the joint precipita- 
tion of brown oxide can be easily checked by stirring. When the 
point IB reached where the oxide only redissolves slowly and no longer 
gives place to a yellow one of hyponitrite, no more silver solution 
should be added. If much more were added, there would be permanent 
precipitation of silver oxide, which is apt to be accompanied by 
silver sulphite. The quantity of silver sulphate required may be 
as much us 40 grams, which means 7 or 8 litres of solution ; 
if silver nitrate be used, about 44 grams will be wanted, dissolved in 
4 litres, or more, of water. 

Half-an-hour after precipitation, the solution is to be poured off, 
even though still a little turbid, and the precipitate washed by 
decantation, for there is a very slow deposition of a mirror of 
metallic silver from the sulphite solution, which goes on for days. In 
order to separate the hyponitrite from the metallic silver and its 
oxide, and perhaps chloride, it has to be dissolved in dilute acid and 
reprecipitated. If every trace of nitrite is to be kept out of the 
hyponitrite, nitric acid can hardly be used, because I find that it 
always contains some nitrous acid, and. it is, therefore, necessary to 
use sulphuric acid. Since the hyponitrite must be kept in solution as 
short a time as possible, it is advisable to have the acid not very 

VOU LXXV. Digitized by Gdbgle 


dilute, in order to reduce the volume of liquid to be filtered. But 
high dilution is better, because the stability of hjponitrite falls 
off rapidly with increasing concentration; moreover, if the sul- 
phuric acid is not dilute enough, silver sulphate will separate; 
a 1 per cent, solution of the acid, well cooled in ice, is suitable, 
some 5 litres of it being required. The precipitate should be treated 
with the acid in portions at a time, not all together; and, as far as 
possible, the undissolved precipitate should be kept off the filter until 
the last. For so long a filtration a Lunge filter-tube-extension of the 
funnel is more suitable than the filter pump, the filtrate being allowed 
to fall directly into excess of sodium carbonate solution. Working 
with these precautions, the silver hyponitrite can be dissolved and 
reprecipitated, even in hot weather, with hardly any appreciable loss. 

Having replaced the alkaline mother liquor by water, dilute sul- 
phuric acid is cautiously added, until, after stirring up the precipitate 
well, the solution is no longer alkaline, and some of it, when filtered, 
is found to contain a trace of the silver hyponitrite ; this is best ascer- 
tained by adding one or two drops of sodium carbonate solution to 
about 100 C.C. of it, which should cause a permanent, yellow, very 
slight, opalescence. 

The precipitate, thoroughly washed by decantation and dried on a 
filter at the ordinary temperature in a desiccator, in the dark, and 
then at 100°, will give 78 per cent, silver (calc. 78-26). But in order 
to preserve the bright colour of the salt and its entire freedom from 
nitrite, all work on it should be done with very little exposure to 
bright daylight. The weight of silver hyponitrite obtained from the 
quantity of sodium nitrite employed should be not less than 
17 grams. 

Imperial Tokyo Univeesity, Japan. 

XI. — Absorption of Nitiic Oxide in Gas Analysis. 

By Edward Divers. 

It is well known that the vapour tension of nitric oxide, dissolved in 
the solution of a ferrous salt, interferes with the use of this reagent 
to remove nitric oxide from other gases. There is, however, another 
absorbent for nitric oxide which leaves nothing to be desired, whose 
use and value have remained unknown. This is a strong solution of 
either sodium or potassium sulphite to which a little alkali hydroxide 
has been added. It quickly absorbs every trace of nitric oxide, which 
it fixes in the form of hyponitrososulphate, NajN^OjSO^. I have 

_. , ^oogle 


already made satisfactory use o^ it to analyse the mixture of nitric 
oxide and nitrogen which is left on heating silver hyponitrite and 
allowing the solid and gaseous products to cool in contact with each 
other. The sulphite need not be very pure, the presence of sulphate 
or carbonate being of no importance. If carbon dioxide or other acid 
gas is present along with the nitric oxide, it is removed by alkali 
before using the sulphite mixture. 

XII. — Interaction of Nitric Oxide with Silver Nitrate. 

By Edward Divers. 

Having reason to think that silver nitrate might interact with nitric 
oxide if heated in it, and there being no information obtainable on 
this point, I have made some experiments on the action of nitric oxide 
on silver nitrate, as well as on other nitrates. 

In the first place, something had to be ascertained as to the be- 
haviour of silver nitrate when heated alone. Heated for 15 minutes 
in dry air or carbon dioxide, it suffers no chemical change until the 
temperature is close to the melting point of sulphur (444^), and the 
slight decomposition which occurs at that temperature, being accom- 
panied by an action on the glass, may be due to that action. A 
minute quantity of oxygen seems to be liberated, and there is a very 
alight greying of the faintly yellow liquid ; on cooling and dissolving, 
there is slight turbidity from the presence of silver, and a trace of 
nitrite can be detected. Only at a much higher temperature does the 
Bait decompose with free effervescence, and then nitric peroxide 
accompanies the oxygen, and silver is deposited ; even then, nitrite 
is present in the mass only in very small quantity at any time, there 
never being enough to remain undissolved when the nitrate is treated 
with a little water. This is sufficient, however, to show that the pri- 
mary decomposition of silver nitrate by heat alone is into silver nitrite 
and oxygen, the instability of silver nitrite at much lower tempera- 
tures, although diminished by the presence of nitrate (Trans., 1871, 
24, 85), accounting fully for its being found in such small quantity 
when the temperature is high, and for the production of nitric peroxide 
and silver instead. As determined by Carnelley, the melting point of 
silver nitrate is 217''. 

The nitric oxide used for the experiments was prepared by the 
ferrous sulphate method, stored for 2 days in a glass gas-holder, and 
dried in its passage to the silver nitrate by a sulphuric acid column. At 
starting, the air in the drying apparatus and in the tube containing 

_. pJoogle 


the silver nitrate was expelled by carbon dioxide, the silver nitrate 
being heated in the gas, in order to dry it. Interaction between the 
silver nitrate and the nitric oxide was recognised by the reddening of 
, the gas, and at the end of an experiment the gases were expelled by 
carbon dioxide before opening the tube. 

Silver nitrate, when heated in nitric oxide, is strongly affected by 
it, being freely decomposed at a much lower temperature than that 
at which it decomposes when heated alone, the nitric oxide becoming 
oxidised. The action commences, perhaps, at 150^ but this is doubtful ; 
at the melting point of the salt, it becomes marked, and at the boiling 
point of mercury considerable, but even at this temperature it is a 
long time before the decomposition is complete, the progress of the 
change gradually becoming slower. For some time, the products are 
silver nitrite and nitric peroxide, AgNOg + NO = AgNOg + NOj, but 
very little silver is liberated, the nitrite being almost wholly pre- 
served for a long time, through combination with the undecomposed 
nitrate. But when, as the nitrate becomes decomposed, the nitrite 
is no longer unprotected, it suffers decomposition, as usual, into 
silver and nitric peroxide ; finally, nothing but silver remains. 

Theoretically, it is quite probable that nitric oxide does not, after 
all, act directly on silver nitrate. For, making the allowable suppo- 
sition that, to a minute extent, silver nitrate decomposes into silver 
nitrite and oxygen, at temperatures much below that at which it does 
so sensibly, the nitric oxide ma^ be regarded as being active by com- 
bining with this oxygen, and thus, by removing it, greatly hastening 
the spontaneous decomposition of the nitrate. This decomposition, 
thus assisted, and occurring at temperatures at which silver nitrite is 
comparatively stable in presence of nitrate, the nitrite remains, although 
at higher temperatures it decomposes almost as fast as it is formed 
from the nitrate. According to this theory, silver nitrate is not 
actually decomposed by nitric oxide, but only decomposes much more 
rapidly in its presence, in consequence of its interaction with one of 
the products of decomposition. For practical purposes, silver nitrate 
and nitric oxide may, however, be treated as acting on each other 
when heated together. 

Nitric oxide has no action on sodium potassium or barium nitrate, 
even at the temperature of boiling sulphur. 

Lead nitrate soon begins to decompose by heat alone, and nitric 
oxide seems to be without effect on its decomposition. According to 
Stas, lead nitrate begins to decompose somewhere above 200° ; I find 
that, for its fairly free decomposition, a temperature not much below 
400° is required. At the boiling point of sulphur even, the decompo- 
sition proceeds at such a moderate rate that only after 10 minutes 
heating does the salt show distinct signs of fusing. No nitrite is pro- 

„.gitized by Google 


duoed, bat a very little peroxide of lead is formed. By washing the 
mass with cold water, and then boiling it out with water, the beauti- 
ful, crystalline, white salt, Fb(0H)N03, can be obtained in abundance. 

XIII. — Preparation of Pure Alkali Nitrites. 
By Edward Divers. 

When pure sodium or potassium nitrite is wanted, it is customary to 
prepare silver nitrite from crude alkali nitrite, and convert this again 
into alkali nitrite by means of sodium or potassium chloride. The 
crude nitrite must be nearly free from sulphate, and either before or 
after adding the silver nitrate to it, nitric acid must be added until all 
the hydroxide and carbonate are neutralised. The silver nitrite is got 
in the most convenient form for washing by precipitating it from con- 
centrated solutions. Silver chloride is very sensibly soluble in a 
concentrated solution of alkali nitrite, and when the solution is no 
longer clouded by the addition of more alkali chloride, it already 
contains this salt in excess. Therefore, somewhat large dilution is 
necessary, and this entails, of course, much evaporation afterwards. 

The silver nitrite process is evidently not a very satisfactory one, 
and when sodium nitrite is wanted pure, it is better to recrystallise, 
three times over, the commercial 96 per cent, sodium nitrite, well 
draining each time on the suction funnel. A concentrated solution 
of the crude salt should be left to clear from lead turbidity for two 
days, or be filtered cold through a fine filter ; the lead carbonate is 
more soluble in the hot nitrite solution than in the cold. After 
separating the lead, the solution should be fully neutralised with 
nitric acid before evaporation for crystallising. Potassium nitrite is 
too soluble and deliquescent to be conveniently purified in a similar way. 

A most satisfactory and simple process for preparing either sodium 
or potassium nitrite, when the pure hydroxide or carbonate is at 
command, is to saturate this with red nitrous fumes under appropriate 
conditions. That nitrites can be thus obtained is known to every 
chemist, was known to Ckky-Lussac in 1816, and was described by 
Fritzsche in 1840, but it has hitherto been stated and believed that 
much nitrate is then unavoidably formed along with the nitrite. 
That is a mistake, and therefore this note is published. If obvious 
precautions of the simplest kind are taken, so little nitrate, if any, is 
formed as to be hardly detectable with certainty in presence of so 
much nitrite. Consequently, if the quantity of pure alkali taken is 
known, a solution of given strength in nitrite is perhaps better 
prepared in this way than in any other. 

Digitized by VjOOQIC 


Avoiding, so far as practicable, the use of cork and caoutchouc, 
nitrous gases, from nitric acid and starch or arsenious oxide, are 
passed into the concentrated solution of the hydroxide or carbonate 
until the alkali is quite neutralised. Sodium carbonate alone is 
somewhat inconvenient, because of its sparing solubility, but this may be 
circumvented by adding it, finely divided and in sufficient quantity, to 
its own saturated solution, just before passing the gases and by often 
shaking the vessel during their absorption. To prevent free access of 
air, the nitrite is prepared in a flask with its mouth kept loosely 
closed while the gases are passing; it is not necessary to cool the 
flask. The strength of the nitric acid and the temperature of the 
generator of the nitrous gases must be so regulated that just a little 
nitric oxide is in excess of the nitric peroxide, and therefore is passing 
unabsorbed, as a guarantee that the latter does not act on the solution 
in absence of its equivalent of the former and thus produce some 
nitrate. To free the gases from volatilised nitric acid, they may be 
passed through a bottle or tube, either empty or packed loosely with 
cotton. The finished solution must be almost neutral, and if acid 
must be boiled until neutral, before exposing it to the air. A concen- 
trated solution of alkali nitrite dissolves a little nitrous acid without 
decomposing it, as water alone would. To get the salt in the solid 
state or to crystallise out the sodium nitrite where it is necessary to 
be sure of absence of all nitrate in it, the solution may be freely 
evaporated, even at a boiling heat, without decomposing or oxidising it. 
The alkali nitrites have been very imperfectly described, and need 
examination. In the meantime, some points in their description are 
here given. Sodium nitrite and potassium nitrites are distinctly 
though faintly yellow, and give markedly yellow solutions in a little 
water. They are very slightly alkaline to litmus. At 15°, 6 parts of 
sodium nitrite requires 6 parts of water to dissolve it; potassium 
nitrite is soluble in about one-third of its weight of water. Sodium 
nitrite melts at 271° (mercury-thread immersed). Sodium nitrite is 
moderately deliquescent, remaining dry in winter-cold air ; potassium 
nitrite is exceedingly deliquescent, and is obtained in very small, 
thick, prismatic crystals, whilst sodium nitrite crystallises in very 
thin, flattened prisms, often very large. Sodium nitrite is well 
known to be anhydrous ; not so potassium nitrite, crystals of which 
are reputed to contain ^H^O. I have, however, examined small, but 
distinct and separate, crystals taken from the upper part of some 
kilos, of the commercial salt, which had become well drained by long 
standing. They were removed in very dry weather and weighed, and 
then found to lose hardly 1 per cent, on fusion. The anhydrous 
character of the potassium salt was further ascertained by testing a 
cake of minute crystals, prepared by myself, which had been pressed, 

• Digitized by Google 



under cover, between porous tiles, in cold, dry air ; the loss of weight 
on heating, much above 100% was a little over 1 per cent., and the 
percentage of potassium was 45*30, instead of 45*88, required for the 
anhydrous salt. 

Somewhat remarkably, the point as to hydration of potassium 
nitrite was examined independently in the same year, 1863, by Lang 
and by Hampe, with identical results, indicating the composition 
expressed by 2KNO2 + H2O, but both these chemists made the 
determination on a magma of indistinct crystals, which had been 
di'ied in a vacuum desiccator. 

XIV. — Reduction of an Alkali Nitrite by an Alkali Metal. 

By Edward Divers, M.D., D.Sc., F.R.S. 

It is already known what are the products which may result from the 
action of sodium amalgam on a solution of sodium nitrate or nitrite. 
Schonbein (1861) first observed the formation of nitrite by the action 
of metallic sodium on a solution of a nitrate; and de Wilde (1863) 
that nitrous oxide, nitrogen, and ammonia are the products of the 
action of sodium amalgam on a solution of potassium, sodium, or am- 
monium nitrate, or on potassium nitrite ; he found that, except alkali 
hydroxide, nothing else is produced, and, in particular, no hydrogen. 
Some years later (1870), however, it was recognised by Fremy, 
aided by a suggestion of Maumen^'s, that hydroxylamine, or what 
appeared to be hydroxylamine, was a product of the reduction. Then 
came (1871) my own discovery of the hyponitrites, together with the 
observation that alkali nitrates in solution are largely convertible into 
nitrites by sodium amalgam, an extension of Schonbein's experience. 
Lastly, Haga and I (1896) proved that the actively reducing sub- 
stance observed by Fremy is actually hydroxylamine, as it had been 
taken to be by Maumen6 and by him, and not hydrazine, as it might 
have been. By a mistake, already pointed out and corrected by me 
{Annalen^ 1897, 295, 366), the discovery of the hyponitrites has been 
in recent years attributed to Maumen^. It will suffice here to say 
that this veteran French chemist has, it so happens, published, in 
another connection, that he had not experimentally investigated the 
redaction of nitrites in solution, and that^ far from laying claim to the 
discovery of hyponitrites, he at first denied its truth on theoretical 
gronnds (Trans., 1872, 25, 772 ; Chem. News, 25, 153 and 285). 

Nitrons oxide, nitrogen, hydroxylamine, ammonia, sodium hypo- 
nitrite, and sodium hydroxide (from the nitrite as well as from the 

Digitized by VjOOQIC 

88 divers: reduction of an 

metal), are, according to my experience, always produced in the re- 
duction of sodium nitrite or nitrate by sodium amalgam, but in pro- 
portions which vary greatly within well-marked limits. Nearly 
one-sixth of the nitrogen can be obtained as sodium hyponitrite in 
one way of working, or scarcely any at all in another. So, too, the 
range of production of hydroxylamine is from nearly 9 per cent, of the 
nitrogen of the nitrite down to one-third per cent. The presence of 
ammonia may be very strongly manifest, or be hardly perceptible and 
escape notice. The two gases, nitrous oxide and nitrogen, together 
represent at least 80 per cent, of the total nitrogen, and may vary in 
relative proportion to the extent of either of them being nearly 
absent. Necessarily, all the sodium of the nitrate or nitrite, not left 
as hyponitrite, appears as hydroxide, along with that derived from 
the metallic sodium used as the reducing agent. So long as any 
nitrite remains, hydrogen does not occur among the products, unless 
a very large quantity of water is present, whilst, if there is very 
little water, hydrogen is not evolved even after all the nitrite is 

Within the limits indicated, the proportioning of the products of 
the reduction is well under control. The conoentration of the solu- 
tion of nitrite, or, to put it better, the relative quantity of water 
present, exercises most influence, the only other circumstance affect- 
ing the course of the reduction being the temperature. The concen- 
tration of the sodium in the amalgam and the proportions of the 
sodium and the nitrite have no direct effect on the reduction. Work- 
ing with a sufficiently concentrated solution of nitrite, the proportions 
of the products remain constant throughout the reduction of the 
nitrite. Probably this is the case also when an exceedingly dilute 
solution is used ; but with a somewhat dilute solution, say 1 in 30, 
there is some difference, due to the fact that the presence or absence 
of much sodium hydroxide modifies the proportions of the products, 
and that this substance is generated so largely. A dilute solution of 
sodium nitrite may be made to behave as a concentrated solution in 
the mode of its reduction by nearly saturating it with sodium 
hydroxide before bringing it in contact with the sodium ama^;am ; but 
the addition of sodium hydroxide to a concentrated solution of nitrite 
before reducing it by the amalgam has no sensible effect, for the 
reason, no doubt, that, in the reduction of the nitrite as it actually 
occurs, about 3^ mols. of sodium hydroxide are produced for 1 mol. of 
nitrite reduced, quite enough, therefore, of itself to make the water of 
a concentrated solution almost proof against the action of sodium. 
With a large quantity of water present, the sodium hydroxide formed 
is not enough to render the water inactive, and in this is to be found 
the explanation of the great difference observed in the proportions of 

„ .gitized by VjOOQ _ _ 


the products, according as the nitrite is dissolved in much or little 

In order to produce as much kyponitrite as possible, little more is 
necessary than to work with a concentrated solution of the nitrite 
(1 of sodium nitrite to 3 or 3^ water), to add the amalgam in some 
excess, and not to allow the temperature to rise above 100^. To 
get as much kydroxylamine as possible, the solution of nitrite must 
be dilute (say 1 in 60), and be kept cold during the addition of the 
amalgam. To preserve the hydroxy lamine from reduction to ammonia, 
the solution must be kept well agitated over the amalgam, and be 
poured off from it as soon as nearly all the nitrite has been reduced. 
Much more time is needed to reduce a dilute solution than a concen- 
trated one. The best conditions for producing much hydroxylamine 
do not allow of much more than half the maximum yield of hypo- 
nitrite being obtained at the same time. 

To get much nitrous oxidej the temperature of the solution must be 
kept as low as possible, whilst to get much nUrogenf the temperature 
must be kept high, the strength of the solution of nitrite being 
without effect. The reduction of a very dilute solution of sodium 
nitijte kept very cold is attended with very little effervescence, 
because the quantity of nitrogen produced is very small and the nitrous 
oxide remains dissolved, although it is readily evolved on warming. 
De Wilde has determined the proportions of the gases to each other, 
but only when the nitrite (or nitrate) was in excess of the sodium ; that 
however, is sufficient, since qualitative examination of the gases has 
shown me that variations in the proportions of salt and metal are 
without sensible influence on the composition of the gases, and also 
that this remains apparently unchanged during the progress of the 
reduction if the temperature is kept tolerably uniform. De Wilde 
found that dilute solutions of nitrite or nitrate of sodium or potassium 
gave larger quantities of nitrous oxide in proportion to nitrogen when 
the solutions were dilute than when they were concentrated, from 
which it might seem that the strength of the solution does affect the 
proportions of the gases to each other; but in the experiments 
conducted by de Wilde, the much greater rise of temperature when 
concentrated solutions are acted on fully accounts for the results he 

Ammonia can always be detected from the beginning of the reduc- 
tion (Thum thought not), but its amount may be minute throughout. 
It can be got in considerable quantity by using a cold dilute solution, 
as for producing hydroxylamine, and, after the main action is over, 
shaking it with amalgam in a stoppered bottle until all the hydroxyl* 
amine has disappeared. It can also be got somewhat concentrated 
for a short time by dropping the concentrated solution of the nitrite on 

„.gitized by Google 


to much solid sodium amalgam, as was first observed by de Wilde 
but even then much hyponitrite is produced. Very hot and dilute 
solutions of nitrite treated with sodium amalgam give little else than 
ammonia and nitrogen. 

The reduction of potassium nitrite by potassium amalgam closely 
resembles that of the sodium salt by sodium amalgam, in every respect, 
both quantitative and qualitative. 

If, for the moment, nitrogen and hydrozylamine be disregarded, as 
they well may be, since their proportions become very small under 
suitable circumstances, the nitrite may then be said to be reduced 
simply to hyponitrite and much of this hydrolysed into nitrous 
oxide and sodium hydroxide. This, at one time, I, as well as other 
chemists, supposed to be the case. But, for a long time now, I have 
felt that most of the nitrous oxide and sodium hydroxide must have 
another origin. Thum has expressed himself in the same sense, 
basing his opinion upon the comparative stability of sodium hyponitrite 
in strongly alkaline solution, for it is only gradually decomposed even 
when boiled with it. This fact by itself, however, is not inconsistent 
with the assumption that the nitrous oxide and sodium hydroxide 
represent decomposed hyponitrite. But it does not stand alone ; for 
(a) Hot concentrated solutions of nitrite yield quite as much 
hyponitrite as cold ones, unless the temperature is well above 100°, 
and even then the yield does not fall off much. (6) In all cases, the 
effervescence accompanying the formation of hyponitrite goes on 
exclusively at the surface of contact with the amalgam, (c) Low 
production of hyponitrite is not attended with higher production of 
nitrous oxide. All these facts are opposed to the view that the nitrite 
is all reduced to hyponitrite in the first place ; so, too, is what follows. 
Although the proportions of the products of the reduction of the 
nitrite vary greatly with the circumstances, it is only within well- 
marked limits J thus, of the nitrite reduced there is from a sixth , 
under one set of conditions, to almost a fifth, under other conditions, 
which becomes partly hyponitrite and partly hydroxylamine (and 
ammonia), whilst the rest becomes nitrogen and nitrous oxide, one or 
the other predominating, according to circumstances. So, too, in one 
extreme case, nearly all of the one-sixth of the nitrite will change 
into hyponitrite, very little becoming hydroxylamine ; or, on the 
other hand, of nearly one-fifth of the nitrite more than half may be 
converted into hydroxylamine, only the rest of the fifth becoming 
hyponitrite. It may, therefore, safely be assumed that about one- 
fifth of the nitrite tends to, or is able to, become hyponitrite, 
although barely one-sixth of the nitrate can yet be secured as this 
salt, because either some of this fifth becomes hydroxylamine, or 
else a little of the hyponitrite is hydrolysed at once or during the 

Digitized by VjOOQIC 


process of isolating it. With that assumption to give more precision 
to the statement, it may be affirmed that many experiments under 
varied ^conditions have shown that about a fifth of the nitrite is 
decomposed by sodium amalgam in one way, and four-fifths in 
another way ; in the one, hyponitrite, hydroxy lamine, and alkali 
(with a very little ammonia and nitrous oxide as secondary 
products) are formed, and in the other, nitrogen, nitrous oxide, and 
alkali; so that when much hydroxylamine is formed it is at the 
expense of hyponitrite only, and when much nitrogen is produced it 
is at the expense of the nitrous oxide only. 

But although this is the case, the hydroxylamine does not seem to 
be derived from the hyponitrite, or the nitrogen from the nitrous 
oxide, but, rather, the one pair of substances is derived' from one 
transition product, and the other pair from another transition 
product. It was pointed out in my first paper that sodium amalgam 
does not act on hyponitrite, and this has since been more fully 
established by Dunstan and Dymond, and again by Thum ; according 
to the last-named chemist, hyponitrous acid is not reduced even by zinc 
and boiling dilute sulphuric acid. In confirmation of my earlier state- 
ment^ I can now assert that sodium amalgam has no action whatever 
on a solution of sodium hyponitrite saturated with sodium hydroxide, 
even at 80° (and, no doubt, at higher temperatures), and when in con- 
tact with it for days together ; no hydroxylamine, ammonia, nitrogen^ 
or hydrogen is produced. In weaker alkaline solutions, hydrogen is very 
slowly evolved, but still without the hyponitrite being affected. Weak 
alkaline solutions of sodium hyponitrite, however, slowly decompose 
of themselves, and then some of the nitrous oxide may possibly get 
reduced by the sodium amalgam. 

As for the nitrogen, it is evident that only while nitrous oxide 
remains in solution and comes in contact with the amalgam can it be 
reduced, even if it is then (see p. 95). Yet, in order to get much 
nitrogen in place of nitrous oxide, it is necessary to work with hot 
solutions, when the solubility of nitrous oxide is at its lowest. It is 
not essential that the quantity of nitrite should be small in propor- 
tion to the sodium, temperature alone appears to be the condition 
determining the formation of nitrogen in place of nitrous oxide. In 
other words, weak solutions of nitrite and excess of amalgam in no 
degree favour the production of nitrogen rather than of nitrous 
oxide, and the proportion of nitrogen is not greater in the gases 
escaping towards the end of a reduction than at the beginning. 

YeTy different is it with ammonia, which is truly a product of the 
reduction of hydroxylamine (in non-acid solution), and the formation 
of which takes place principally during the final action of the 
amalgam. Against the notion, highly improbable as it is, that the 

Digitized by VjOOQIC 


nitrogen may come from yet undecomposed nitrite and already formed 
ammonia, which would also account for the comparative absence of 
ammonia in the earlier part of the reduction, there may be adduced 
de Wilde's observation, that ammonium nitrate, when reduced by 
sodium amalgam, gives much more nitrous oxide in proportion to 
nitrogen than potassium or sodium nitrate does, no doubt because 
there is less rise in temperature. 

Without speculating on the constitution of a nitrite, we are able to 
see from the interactions between ethylic iodide and silver nitrite 
that a nitrite may react both as an oxylic salt, NaONO, and as a 
halide, NaNOj. From the reduction by sodium there will then first 
result the radicles NaON= and NaNO ; from the former, or sodox- 
imido-radicle, may well come the hyponitrite and hydroxylamine, and 
from the other, or sodium nitroside radicle, the nitrous oxide and 
nitrogen. In accordance with the facts observed, the sodoximide, in 
concentrated alkaline solution, will condense to sodium hyponitrite, 
stable against reduction, or, in very dilute alkaline solution, will, by 
hydrolysis and reduction, become alkali and hydroxylamine. The 
hypothetical nitroside will also condense and simply hydrolyse into 
nitrous oxide and alkali, mainly at low temperatures, or will become 
reduced and hydrolysed into nitrogen and alkali, principally at higher 

To establish the points in the reduction of the two nitrites by their 
respective metals, here described, I have made very many experiments, 
usually working on quarter-gram molecules of nitrite. The hypo- 
nitrite obtained was weighed as silver salt. The hydroxylamine was 
estimated by the quantity of metallic silver it yielded, and in this 
way : the black precipitate it causes in silver nitrate solution, in 
presence of alkali, being largely suboxide, was washed with cold 
dilute nitric acid and ammonia alternately, and the residual brownish 
metallic silver weighed and calculated into hydroxylamine by the 
ratio 2Ag:Nn30, experiments (described in the next paragraph) 
with solution of hydroxylamine sulphate of similar dilution and 
alkalinity having shown that this could be done accurately enough. 

The important observation, made by Thum, that hydroxylamine, 
when oxidised by suitably alkalised mercuric oxide, silT^r oxide, or 
cupric hydroxide, will yield a little hyponitrite and nitrite, induced me 
to ascertain whether, in my experiments, the destruction of hydroxyl- 
amine in this way, sometimes in considerable quantity, might not 
account for some of the hyponitrite afterwards found to be present. 
To ascertain whether this took place, I made a blank experiment very 
similar to those made in studying the reduction of sodium nitrite^ 
except that sodium hyponitrite itself was absent. Thus, hydroxyl- 

Digitized by VjOOQIC 


Amine sulphate, 1*5 grams (^0*6 gram hydrozylamine) was dissolyed 
along with 32 grams of sodium hydroxide in nearly 2 litres of water, and 
then a sedation was ran in, with stirring, of 7*6 grams of silyer nitrate, 
which was a considerable excess, sach as was ased in the other experi- 
ments. The abandant, black precipitate was washed, exhaosted with 
ice cold, dilate nitric acid, and the eolation, neatralised as asaal in my 
other experiments, gave no silver reaction for sUver hyponitrite, and 
nothing more than a slow and yery slight action on permanganic acid, 
which might be dae to a trace of either nitrons or hyponitroas acid. 
It was easily seen that some nitrons acid was formed, by applying the 
iodide and starch test. Under the circamstances of my experiments, 
therefore, even when 7 per cent, of the nitrite had been reduced to 
hydroi^lamine, there coald have been no perceptible prodaction of 
hyponitrite daring the after oxidation of the hydroxylamine. The 
metallic silver, washed out with dilate nitric acid * and ammonia, 
weighed 3*8 grams, the calculated quantity being 3*95 grams. The 
nitrite detected in the mother liquor of the black precipitate had 
been formed in too small a quantity to materially affect the weight of 
the metallic silver. 

Generally, sodium hydroxide was approximately estimated, after all 
the hyponitrite had been precipitated, by titration Of the mother liquor 
with nitric acid, and of the silver oxide that had been precipitated along 
with the silver hyponitrite and the metallic silver. The amalgam used 
was of approximately known strength, ascertained, not by sampling, 
which is impracticable, but by uniformly preparing successive quanti- 
ties, and sacrificing one to assay by dilute sulphuric acid and weighing 
the sodium as sulphate ; after use in reducing nitrite, the sodium re- 
maining in the mercury was sometimes determined in the same way. 
Nitrous oxide and nitrogen were not measured ; their total nitrogen 
was foand by difference, and their proportions had been sufficiently 
ascertained by de WUde, as I have already said ; but their relative 
abundance was estimated by a burning splint of wood, the reduction 
of the nitrite being always conducted in a loosely closed flask. The 
range of this reaction was from that of a gas utterly extinguishing 
eombostion to that of one which supported it most vividly ; in any 
miifonnly conducted experiment, the gases evolved towards the end 
behaved like those given off at first. 
To ascertain the effect on its reduction by sodium of adding sodium 

* It WM proyed many yean ago that silver is inaolable in dilate nitric acid, the 
pfeienoe of nitrons acid being necessary to make it dissolye. Bat the contrary has 
been since asserted to be trae where the silyer is finely divided, as when precipi- 
tated. This error, as 1 most regard it, is due to precipitated silver when black or 
M^>vf^i> containing snboxide, which gives it its colour : this b resolved by acids 
into oride of nlver, which dissolves, and metallic silver, which is left. 

^o^ "^^ Digitized by C^ogle 


hydroxide to a concentrated solution of sodium nitrite (negative as 
this proved to be), two methods were adopted. In one, the amalgam 
was covered with a cold saturated solution of sodium hydroxide) 
which has no action on it, and then the concentrated solution of nitrite 
was slowly added ; at first, the alkali greatly impedes the action of 
the amalgam on the nitrite, but when more of the solution of the 
latter is added, the action goes on faster and to the end, and gives the 
usual large proportion of hyponitrite with very little hydroxylamine. 
In the other, a concentrated solution of sodium nitrite and sodium 
hydroxide was treated with some of the amalgam ; then more sodium 
nitrite was added, and then more amalgam. The result was the same 
as before. The object of working in this way was to obtain the 
effect, if any, of the most concentrated alkali from the firsts without 
having to deal afterwards with a very large excess of alkali when the 
analysis had to be made. 

I have also tried to ascertain the effect of diminishing the amount 
of alkali present. In acid reducing mixtures, nitrous acid becomes 
largely converted into hydroxylamine without production of hypo* 
nitrous acid, so that it seems probable that, could the alkali formed in 
the reduction of the nitrite by sodium be neutralised nearly as fast as 
it is produced, much hydroxylamine would be obtained and very little 
hyponitrite. The use of the ordinary acids for the purpose in such a 
way as to give conclusive evidence on the point does not seem to be 
practicable, whilst «the great rapidity of the process of reduction 
makes the use of carbon dioxide (Aschan, ^ar., 1891, 24^ 1865) very 
unpromising. I have, therefore, tried the effect of adding ammonium 
acid carbonate along with the sodium nitrite, expecting the ammonia 
to be inactive. In one case, where I used the amalgam in large 
excess, much ammonium amalgam was formed, and, what was quite 
unexpected, neither hyponitrite nor hydroxylamine. In another 
experiment, in which the nitrite was kept in excess of the amalgam, 
the previous addition of the ammonium carbonate in excess was 
without any effect ; the nitrite solution had to be used slightly dilute 
because of the carbonate, and gave, therefore, a little less hyponitrite 
(about 12*7 per cent, of the nitrite consumed) and a little more 
hydroxylamine (about 3 per cent.) than in the best way of working for 
hyponitrite. The presence of the ammonium carbonate was, therefore, 
without effect, the reaction between the nitrite and the sodium being 
already complete when the sodium oxide comes in contact with the 
water and ammonium carbonate. 

I satisfied myself that a fairly concentrated solution of nitrite is 
uniformly reduced from the commencement to the end of the reaction 
if the temperature is kept tolerably constant, the method employed 
being to examine the gases in the way described, and the hyponitrite 

Digitized by VjOOQIC 


Udd bydrozylamiiKe as follows. To a solution of nitrite, one-half only 
of the quantity of sodium amalgam required to reduce it was added, 
and it was then found to contain hyponitrite and hydrozylamine in 
the same relative proportion as if the nitrite had been f uUy reduced 
(with cooling), and in apfHrozimately half the quantities the nitrite 
would have yielded if the f uU amount of sodium amalgam had been 

Sodium amalgam was proved to have little or no action on nitrous 
oxide by exposing the gas for a long time to its action. The amalgam 
was liquid, and, when shaken up with the moist nitrous oxide in a 
stoppered bottle, coated the sides of the bottle. With occasional 
vigorous shaking, the bottle was kept closed four days ; when opened, 
it was found to contain the nitrous oxide little, if at all, deteriorated 
as a supporter of combustion. In another similar experiment, a 
saturated solution of sodium hydroxide was poured over the amalgam ; 
in this case, the amalgam did not coat the sides of the bottle, but the 
solution served to keep the dissolved nitrous oxide in contact with the 
amalgam. The bottle was often vigorously shaken, and was not 
opened until after four days. The nitrous oxide was almost or quite 
unchanged. Holt and Sims have studied the oxidation of soditmi and 
potassium by nitrous oxide, but only at much highcor temperatures 
than those in these experiments, which were at 26 — 30^. 

Imfbrial Tokyo Uhiyebsitt, Japak. 

XV. — Hyponitrites ; their Properties^ and their Pre- 
paration by Sodium or Potassium. 

By Edward Diybbs, M.D., D.Sc., F.R.8. 

Thb hyponitrites have received the attention of many chemists 
besides myself since their discovery in 1871, and even this year new 
ways of forming them and the new working of an old method have 
been published. Yet much has been left to be put on record before a 
fairly correct and full history of these salts can be said to have been 
given, and the present paper is meant to be the necessary supplement 
to what has already been published. 

Ways qf forming Hyp<mi4TiU8, 

Ko writer on hyponitrites in recent years has shown himself 
acquainted with all the known ways of getting these salts, <x even 
with the most productive. The following complete list is valuable, 

Digitized by VjOOQlC 


therefore^ and is of special interest as bringing together the various 
modes o? formation of these salts: 

1. Reduction of an alkali nitrite by the amalgam of its metal 
(Divers, 1871). 

2. Beduction of an alkali nitrite by ferrous hydroxide (Zom, 1882 ; 
Dunstan and Dymond). 

3. Reduction of (hypo)nitrososalphates by sodium amalgam (Divers 
and Haga, 1886). 

4. Reduction of nitric oxide by alkali stannite (Divers and Haga, 

5. Reduction of nitric oxide by ferrous hydroxide (Dunstan and 
Dymond, 1887). 

6. Decomposition of a hydroxyamidosulphonate by alkali (Divers 
and Haga, 1889). 

7. Oxidation of hydroxylamine by sodium hypobromite (Kolotow, 

8. Oxidation of hydroxylamine by mercuric oxide, silver oxide, or 
cupric hydroxide (Thum, 1893). 

9. Interaction of hydroxylamine and nitrous add (Thum, H. Wis- 
lioenus, Paal and Elretschmer, Tanatar, 1893). 

10. Oxidation of hydroxylamine by bensenesulphonic chloride and 
alkali (Piloty, 1896). 

11. Interaction^ in methylic alcohol, of hydroxylamine with nitrous 
gases (Ejiufmann, 1898, Annalen, 200, 98). 

12. Interaction, in methylic alcohol, of hydroxycarbamide and 
nitrous gases (Hantzsch, 1898). 

13. Interaction of dimethylhydroxynitrosocarbamide and alkali 
(Hantzsch and Sauer, 1898). 

Menke's reduction of fused alkali nitrate by iron, and R&y's reduc- 
tion of mercuric nitrite by potassium cyanide in solution, are not 
included in the list, because both reductions are very doubtful, and 
require confirmation before they can be accepted. In the present 
paper, only the original method of preparing hyponitrites will be 
treated of. 

Preparation of Sodivm HyponUriU SohUicn by the Reduction of 
Sodium Nitrite with Sodium Amalgam. 

Sodium nitrite can be converted by sodium amalgam in the easiest 
and quickest imaginable way into fully one-sixth of its equivalent of 
sodium hyponitrite; this remains in solution, and is pure but for 
the presence of much sodium hydroxide. From this solution, the 
sodium salt itself, as well as silver hyponitrite, can be at once pre- 
pared, nearly pure and with hardly any loss. The solution, if 

Digitized by VjOOQIC 


caatioosly neatralisedy is also at once fit for preparing lead, copper, 
mercury, and some other salts. The neutralisation is known to be 
complete when a little ot the solution just ceases to give black oxide 
when mixed with a drop of a dilute solution of mercurous nitrate. 
Others who have tried this method, and particularly Hantzsch and 
Eaufmann, got far less favourable results. 

Pure sodium nitrite is necessary, but that can now be prepared 
very simply (this vol., p. 85). In order to get as much hyponitrite 
as possible and as little hydroxylamine, the nitrite must be in con- 
centrated solution ; three times its weight of water seems to be the 
best quantity to dissolve it in when operating in the way to be 
deecribed. Using these proportions, there is enough water to form, 
with the sodium oxide produced, a solution of the composition 
NaOH+3HsO, which is a nearly saturated solution of NaOH,H30 at 
the mean temperatura^ In presence of so much hydroxide, the 
water is also quite saturated with hyponitrite, a small quantity of 
this salt even separating when the solution is kept at 0^ for a 

To reduce sodium nitrite in cold concentrated solution, 2| atoms of 
sodium are needed, the additional half atom being consumed in the 
unavoidable formation of some nitrogen, hydroxylamine, and 
ammonia. This accords well enough with the statement in my first 
paper, as a first approximation, that not more than 4 atoms are 
active on sodium nUraie, In practice, however, 3 atoms of sodium 
should be used in reducing sodium nitrite, partly because it is wanted 
afterwards to reduce hydroxylamine, and partly because it is im- 
pmrtant that all the nitrite should be reduced, and this, notwith- 
standing statements to the contrary, can only be accomplished quickly 
in presence of a good excess of sodium. The strength of the amalgam 
18 not an essential point ; I have, however, found it most con- 
venient to work with a soft, solid amalgam having the composition 
(NaHgg),, or 23 grams of sodium to 1600 grams of mercury. t The 
temperature, also, is not of importance if only the solution of nitrite 
is concentrated, and although it may in fact rise nearly to 100^ with- 
out harm, it is better to follow my original direction to keep the 
flask in a stream 6i cold water during the reduction. It is, however, 
preferable to cool it, particularly in warm weather, by means of a 
brine and ice bath, as then the amalgam can be added much faster 
without producing any great evaporation. The temperature of the 
Bidntion during the reduction then ranges, with a convenient rate of 

* Sodium hydroxide formB a saturated solation at 16^ in its own weight of water. 
When coolod, this solntion deposits large pointed prisms of the monhydrate. 

t Taaatar erred in supposing that I recommended the use of hard amalgam, 
and his supp ose d improvement of my process is not one in hct 

Digitized by VjOOQIC 


working, from 5° to 25^ and the time taken to add 23 grams of 
sodium need not be more than 10 minutes. 

From a quarter to a half gram-molecule i>f sodium nitrite is a oon- 
▼enient quantity to work on, and the solution is best contained in a 
360 to 460 C.C. pear-shaped, wide mouthed flask, lying very obliquely 
in the cooling bath while the amalgam is added by means of a 
spatula. The last fourth of the amalgam may be put into the flask 
as rapidly as it can be, and the flask may then be removed from the 
bath. It is kept actively rotated for 10 — 16 minutes, during which 
the temperature will rise to about 40° and then f alL The solution 
and mercury are next poured into a narrow mouthed stoppered bottle 
so as to half fill it, the thick, aqueous solution adhering to the flask 
being washed into the bottle, but the water used should be limited to 
2 or 3 C.C. if it is desired to obtain the solid sodium salt. The whole 
is now violently shaken for 10 minutes or so, so as to destroy all the 
hydrozylamine. To ascertain this, a drop of the solution is tested by 
diluting it and adding a drop of silver nitrate solution, and a slight 
excess of dilute nitric acid ; there should not be the slightest black 
tint due to silver reduced by hydrozylamine. No gas is liberated 
during the shaking, but a very strong odour of ammonia is developed. 
Strange to say, a minute quantity of nitrite is still present, and it 
seems almost impossible to entirely remove it, although it can be so 
far reduced by an hour's shaking of the solution with the amalgam 
that the acidified solution does not blue potassium iodide and stcuroh 
until it has stood for about an hour. 

On separating the solution from the amalgam and exposing it in 
a dish overnight over sulphuric acid, under reduced pressure, it will 
be free from ammonia, and is virtually a pure and stable concen- 
trated solution of sodium hyponitrite and hydroxide. 

As here described, the preparation of a solution of sodium hypo- 
nitrite ready for use is the same as that followed by me in 1871 (with 
nitrate), except the important modification in the manner of removing 
the hydroxylamine. When silver hyponitrite is prepared from the 
crude solution, the hydroixylMsiine gets destroyed by silver oxide^ as 
I pointed out in the addendum to my first paper. Zom, as an 
improvement, introduced the use of mercuric oxide, on the ground 
that destruction of some silver hyponitrite was thus avoided, but 
he overlooked the fact that it is the silver oxide, just as it is mer- 
curic oxide^ which becomes decomposed, the hyponitrite or any other 
add radide being untouched by the hydroxylamine in alkaline solu- 
tion. Whether, therefore, mercuric oxide, or silver nitrate, or mer- 
curic nitrate is used, and the precipitated metal then separated, the 
result is just the same in concentrated alkaline solutions, except that 
the dropping in a solution of the nitrate is more easy to carry out 

Digitized by VjOOQIC 


thftn stirring ap with merouric oxide. Farther, where the alkaline 
soliiiion is very weak, the use of mercnry compounds ia not without 
objeetiony as a little merourio oxide remains in solution. But whether 
silver or mercury oxide is employed, the result ia unsatisfactory, for, 
as Thnm has pointed out, both these oxideSy in destroying the hydr- 
oxylamine, regenerate nitrite. Not, however, that Thum himself 
found this prevented him from successfully purifying the silver hypo- 
nitrite from nitrite by thorough washing and reprecipitation. Ber- 
thelot and Ogier, Paal and Kretschmer, and I myself, have not, 
however, met with the same success, as I found it necessary, in order 
to get silver hyponitrite free from all trace of nitrite, to begin by 
pneiidtating it in the absence of nitrite. Nevertheless, far from 
doubt on Thum's statement, I believe his silver salt to have 
me of the purest ever prepared, from the account he has given 
of the properties of hyponitrous acid. No .one, however, will be 
di^Msed to deny the superiority of sodium as a means of removing 
the hydroxylamine from the solution. 

An almost pure solution of sodium hyponitrite can be conveniently 
got by dissolving the freshly prepared, hydrated, solid salt in water. 
Sodium iodide, or potassium iodide and the silver salt, will also furnish 
a solution of alkali hyponitrite. To get a solution for precipitating 
purposes, Thum proceeded in an indirect way, first preparing a solu- 
tkm 6t hyponitrous add, and then adding enough sodium hydroxide to 
make the solution neutral to phenolphthalein, an effective but very 
wasteful process. Kirsohner also, wanting a solution for precipitating 
purposes, used sodium chloride and silver hyponitrite, which, in 
complex and wasteful way, he made to yield a solution which although 
mixed with much chloride and nitrate, was practically free from silver. 

Sodium H^panUrite. 

In 1878, Menke gave full analyses of crystals of a stable salt 
having the composition of sodium hyponitrite containing GH^O, 
whieh he had prepared by deflagrating in an iron crucible a mixture 
of sodium nitrate and iron filings, keeping the product at a red heat 
for an hour in a fire <rf charcoal rather than of gas, boiling the mass 
with water, filtering off iron oxide, evaporating, and leaving to 
etystallise. He makes no reference in his paper to the large amount 
of sodium hydroxide he must have had to deal with, although this 
should have seriously affected the procedure. In 1882, Zom sub- 
mitted Menkens method to a thorough examination, but failed to obtain^ 
the least trace of hyponitrite; he found, however, that ferrous 
hydroxide, acting on a solution of sodium nitrite, did produce sodium 
hyponitrite (in solution). His suggestion that Henke bad mistakep 

Digitized by VjOOQIC 


carbonate for hyponitrite takes no account of the fact that tlie 
nitrogen and water in the salt were repeatedly determined. It can 
now, however, be stated with certainty that Menke's salt was not the 
sodium hyponitrite obtainable by reducing sodium nitrite by sodium 
amalgam and water, for this differs from it in degree of hydration, 
stability, and in other properties. 

D. H. Jackson (Froc, 1893) described two ways in which he had 
succeeded in preparing sodium hyponitrite, but with such difficulty that 
he was deterred from investigating its properties ; indeed, he merely 
mentions, in proof of his success, that he obtained crystals which con* 
tained the theoretical proportion of sodium, but this happens to be no 
proof at all, as sodium carbonate has just the same content of sodium ; 
moreover, the hyponitrite is a hydrated salt which cannot be ren- 
dered anhydrous without some decomposition, and although of crystal- 
line texture, the salt can hardly be described as occurring in crystals. 
Nevertheless, his success in getting the salt is not to be doubted* 
One of the methods he adopted was to reduce a concentrated solution 
of sodium nitrate by sodium amalgam, evaporate the solution in a 
vacuum until the salt crystallised, and wash the crystals with alcohol 
to free them from sodium hydroxide. With some modificati<»i, the 
process he followed is an excellent ona 

To obtain sodium hyponitrite from the somewhat thick solution pre- 
prepared as already described, it is first passed through a €k>och asbestos 
filter well covered from the air ; it contains 1 mol. of sodium hydroxide 
to 3 mols. of water, and is a saturated solution of the hydroxide, whUst 
there are about 21^ atoms of sodium present as hydroxide to 1 as 
h3rponitrite. Cooling alone will cause the separation of the hypo- 
nitrite, and the solution readily loses water in a vacuum over sulphuric 
acid until it retains only about 2 mols. to 1 mol. sodium hydroxide, 
when almost the whole of the sodium hyponitrite will have separated ; 
at 25 — 30^, this will happen in about 40 hours, the salt separating as 
minute, crystalline granules, some of which adhere to the walls of the 
dish, but most of them being deposited as a thick crust on the 
surface of the solution. Below 15^, the mother liquor readily deposits 
crystals of the monhydrate of sodium hydroxide, and as evaporation 
is slower in the cold, it is better, for both reasons, to conduct the 
operation in a warm room. 

The only effective way of separating the salt from its viscid mother 
liquor is by the pump and a Gooch crucible unlined with asbestos ; 
draining on a tile is impossible. In the same apparatus, it is washed 
with absolute alcohol and then transferred to a basin and gently 
triturated with fresh portions of alcohol until all the sodium hydroxide 
has been removed. If now drained on a good tile, it is the nearly 
pure hydrated salt, but very unstable, losing both water and nitrous 

Digitized by VjOOQIC 


oxide, and consequently becoming contaminated again with sodium 
hydroxide ; if, however, it is promptly dried in a vacuum desiccator, 
it becomes anhydrous before it has undergone much decomposition, 
and is then quite stable in dry air. The amount of the hydrated 
sodium hyponitrite should be quite one-sixth of the calculated quantity, 
and the mother liquor wUl then be too poor in hyponitrite to be used 
as a source of sUver hyponitrite. 

A modification of the method, which gives an equally good yield, is 
to precipitate the salt by absolute alcohol instead of evaporating ; the 
only precaution necessary is to prevent, as far as possible, the salt from 
becoming attached to the walls of the vessel. A large quantity of 
alcohol is required, because of the very large proportion of sodium 
hydroxide which is present. A few drops of the solution are added to 
the alcohol in a flask and at once violently shaken with it until the 
hyponitrite has completely solidified ; then gradually the rest of the solu- 
tion is poured in with very active agitation. If abundance of alcohol 
is used from the first, with thorough mixing, very little of the salt will 
remain in solution and very little adhere to the flask ; with less 
alcohol at first, a notable quantity of the salt is lost by being kept in 
solution, for although it is afterwards slowly deposited, it is not then 
in a serviceable condition, and much salt is liable to adhere to the 
flask, which can, indeed, be dissolved out in water, and be reprecipitated 
by alcohol, but only with very great loss. 

The action of sodium chloride on silver hyponitrite (see p. 106) is 
complex and quite unsuitable for the preparation of a solution of 
pore sodium hyponitrite. Nevertheless, such a solution, charged as 
it is with sodium chloride, and containing, besides, some silver 
hyponitrite dissolved in it, deposits sodium hyponitrite when mixed 
with much absolute alcohol, and this coDstitutes Jackson's second 
method of getting the salt j it always, however, contains a little 
chloride mixed with it. 

The granular form of sodium hyponitrite is most marked in it when 
it has been separated from a highly concentrated solution of sodium 
hydroxide ; when it is redissolved in a very little water and the solu- 
tion rapidly evaporated in an exhausted desiccator, the salt separates 
as an inmost structureless membrane on the surface, and there readily 
becomes opaque and apparently anhydrous. In the ordinary desiccator, 
a finely granulated crdst forms. I have never obtained it in crystals. 
The Bait, when quite freshly prepared, has an exceedingly mild, alkaline 

The attempts to determine the degree of hydration of the salt have 
been unsatisfactory because of its instability, but they point to the 
formula (NaON),-l-5HsO. That formula requires 23*47 per cent, 
sodium, whflst analysis of the salt weighed as soon as it was almost 

Digitized by VjOOQIC 


free from alcohol, gave 23*66 per cent. In place of 30'61 per cent, for 
the hyponitiite ion, 28*10 per cent, was obtained by dissolving the salt 
in water and precipitating with silver nitrate, a deficiency fairly 
attributable to decomposition before the silver nitrite could be added ; 
for, as proved by Zom, this way of estimating hyponitrous acid is 
accurate. Loss of weight in the vacuum desiccator gave 44*91 per 
cent., whilst the calculated quantity of water is 46*92 per cent., but 
the difference is easily accounted for as due to loss of nitrous oxide, 
and, indeed, would be even greater but for the fact that this loss 
involves fixation of water by sodium oxide. 

The anhydrous salt, somewhat decomposed, is non-coherent and 
opaque, and in appearance much like that of magnuia alba ; heat is 
evolved when it dissolves in water, and it is insoluble in alcohoL The an- 
hydrous salt only slowly takes up water from a solution of sodium hypo- 
nitrite. Heated in a closely-covered vessel, it yields nitrogen and sodium 
oxide mixed with some nitrate, 3(NaON)2 » 2Ns + 2Na,0 + 2NaN0,. 
The salt bears a heat of 300° without decomposing, and then melts and 
effervesces ; glass, platinum, and even silver are freely attacked by the 
fused mass, and the product hisses when water is added to it. Sodium 
hydroxide and nitrite are the solid products when the hydrated salt is 
quickly heated, and nitrous oxide, as well as nitrogen, is given otL 
Strong sulphuric acid decomposes the salt, with production of odour- 
less, white vapours, and does not form nitrosylsulphate if the sodium 
salt is pure. 

The salt, or a fairly concentrated solution of it, effervesces with a 
dilute acid like a carbonate. The solution, with the respective re- 
agents, gives a precipitate of calcium hyponitrite and of most other 
hyponitrites at once. It dissolves a little silver hyponitrite and decom- 
poses silver chloride (see the account of silver hyponitritei p, 106). 
Dry sodium hyponitrite is not decomposed by carbon diaxide» and, 
since the hydrated or dissolved salt partly decomposes by interaction 
with the water, its power of fixing carbon dioxide does not indicate 
that it is directly decomposable by that substance. The solution, 
when boiled, decomposes moderately fast into hydroxide and nitrous 
oxide. If allowed to stand for a day, a trace of nitrite is formed 
(see p. 114). 

PataaHum HyponUriie; PokuHum Amalgam. 

The preparation of a solution of potassium hyponitrite is throughoat 
like that of a solution of the sodium salt. It is only necessary, 
therefore^ to say something concerning the potassium ^^n^ftlgam which 
is required, concerning which as a reagent little or nothing has been 

Digitized by VjOOQIC 


Merely for convenience in working, tlie composition of the potasdiim 
amalgam shonld oorrespond pretty cloflely, in parte by weight, to that 
reecNnmended for the sodium amalgam, namely, (Hgi^K)^, or 2800 of 
mercury to 39 of potassium, this being the weakest anudgam that is 
solid, a pasty amalgam like that of sodium not being obtainable. 
Although it crystallises in simple cubes, often very large, which are so 
sharp angled tiiat they can hardly be introduced into a flask without 
fracturing it, these crystals are very easily crushed, in a porcelain 
mortar, and are then in a state quite convenient for use. Sodium or 
potaflsinm amalgam not stronger than here recommended (1 kilo, of 
mercury to 14 grams of alkali metal) is particularly easy to prepare in 
Draper's way, that is, by melting the sodium or potassium under solid 
paraffin and adding the mercury to it, at first very gradually. The 
operation can be performed on the open table. In spite of the fact 
that more heat is evolved, according to Berthelot's numbers, the action 
is less violent in preparing potassium amalgam than it is in the case 
of sodium amalgam. Potassium, also, nearly always requires to be 
well stirred with a glass rod to bring about its first contact with the 
mercury under the paraffin;* sodium never does. When all the 
mercury has been added, either amalgam requires good stirring in 
order to dissolve all lumps, and should again be stirred when 
solidifying, in order to disturb crystallisation as much as possible. 
The specific gravity of the paraffin is about the same as that of 
potassium, but paraffin expands so very greatly in melting that the po- 
tassium readily sinks in it when it is in the liquid state. Muhlhaeuser, 
many years ago, melted sodium under petroleum and then added the 
mercury to it, and, in recent years, Nef has^ recommended the use of 
toluene, which boils freely by the heat produced in the union of the 
metals. But toluene could hardly be used in making potassium amalgam, 
because of its specific gravity. 

A highly concentrated solution of potassium hyponitrite and 
hydroxide having been prepared, the hyponitrite can be precipitated 
1^ absolute alcohol, but only very incompletely, and some of what is 
precipitated is afterwards dissolved away in washing it with more 
aloolu^ The preparation of this salt is, therefore, less satisfactory 
than that of the sodium salt. 

Another way of making potassium hyponitrite is to decompose sUver 
hyponitrite with exactly the right quantity of solution of potassium 
iodida By rapid evaporation under reduced pressure, the solution can 
be eoncentrated, although with partial decomposition, preparatory to 
treatiBg it with absolute alcohol, and it can even be dried up, so as to 

* This is probably dne to the foot that the potassiom presses but lightly upon 
the mereniy on account of its specific gravity not greatly exceeding that of the 
paraJBn, aad not beeanse of any chemical difference. 

Digitized by VjOOQIC 


yield the impure solid salt. The cold of evaporation in a vacanm has 
sometimes caused the separation from the concentrated solution of 
hydrated crystals, which, however, melt when placed on filter paper ; 
otherwise, the salt is obtained anhydrous in minute, prismatic crystals. 
The salt decomposes more rapidly than the sodium salt, but is stable 
when quite dry. It is soluble in 90 per cent, spirit, and slightly even 
in absolute alcohol. Its aqueous or alcoholic solution yields silver 
hyponitrite with sUver nitrate, dissolves silver hyponitrite to some 
extent, and in other respects behaves like one of sodium hyponitrite. 
It has not been obtained sufficiently undecomposed to be fit for 
quantitative analysis. 

PfBparaiion of Silver ffypamtrite. 

The hyponitrites were discovered through the production of the 
silver salt, and since that discovery this hyponitrite has been prepared 
and redescribed by many chemists ; all deviations from the account I 
first gave of it are, however, incorrect, and the only additional observa- 
tions that have been made are that it can be obtained in a purer state 
than I got it at first, and that it gives off nitric peroxide when heated. 
The very poor success in obtaining it in satisfactory quantity in recent 
years is remarkable (see p. 97) ; this seems to be due to erroneous 
methods of procedure, either in reducing the nitrite or in converting 
the sodium hyponitrite into the silver salt. 

The concentrated solution of sodium hyponitrite and hydroxide, 
already described, is diluted and mixed with just sufficient silver 
sulphate or silver nitrate, dissolved in much water, to precipitate the 
hyponitrite ; for it is unnecessary to neutralise the sodium hydroxide. 
(Neutralisation can, indeed, precede precipitation, if desired, as in pre- 
paring mercury and other hyponitrites, but in the case of the silver salt 
it is quite unnecessary, and there is a risk of loss of hyponitrite.) Silver 
sulphate should be used if it is essential to exclude nitrate from the silver 
hypom'trite ; for, as will be shown, washing and reprecipitating are but 
imperfect means of purifying the precipitate. Supposing a half gram- 
molecule of sodium nitrite to have been reduced, the alkaline solution, 
diluted to three times its volumeor more, is mixed with 13 grams of silver 
sulphate or 14 grams of silvernitrate dissolved in about3 litres of water, 
and the mixture stirred vigorously at once and continuously for 6 
minutes, in order to convert the silver oxide into silver hyponitrite. 
After the precipitate has nearly all subsided, the turbid liquor, if 
bright yellow rather than brownish, is decanted and more silver solution 
added to it until, after stirring well, some brown silver oxide remains, 
when the whole is poured back and stirred up with the main pre- 
cipitate, and then left to settle. Good daylight is almost essential 

Digitized by VjOOQIC 


for jadging the colour of the precipitate when finishing, bat the 
preci{Atate should not be exposed to light more than is absolutely 

After washing somewhat by decantation, the precipitate is stirred 
up with successiye portions of highly dilute sulphuric acid (3 or 4 per 
mille) until this fails to become fully neutralised, and shows, there- 
fore, when poured off and mixed with a drop of sodium carbonate, a 
slight yellow opalescence due to silver hyponitrite. The precipitate, 
after being washed with water by decantation until the washings no 
longer contain sulphate, is stirred up with water containing a trace of 
sodium carbonate, and, finally, again washed with water. It is then 
collected on a filter and dried in the dark in a vacuum. When thus 
dried, it may be heated for a time to 100° in dry air without change 
and become still drier. It is now usually as pure as it is possible to 
get it. As, however, the operations are not always so perfectly 
carried out as to ensure this degree of purity, it is desirable some- 
times to submit the salt to further treatment, preferably before it 
has been dried. In that case, it is dissolved, in portions at a time, 
in 3 per mille ice cold dilute sulphuric acid, and is either ex- 
peditiously filtered, if necessary, into some sodium carbonate solution, 
or, if not, is at once made alkaline with sodium carbonate. The re- 
predpitated salt b then treated with sulphuric acid and washed in 
exactly the same manner as the original precipitate. Even after 
reprecipitation, the silver hyponitrite obtained from 34*5 grams of 
sodium nitrite will weigh about 11 grams. 

Hie process for preparing silver hyponitrite just given differs from 
that contained in my first paper in not neutralising the sodium 
hydroxide with acetic add, in taking silver sulphate instead of 
nitrate, sulphuric acid in place of nitric acid, and sodium carbonate 
in place of ammonia, and in some minor details. The use of sulphuric 
acid is not new, that acid having been first used by van der Flaats, 
but the motive for the change is new and has been already given. 
Cold dilute sulphuric acid is not in the least less active than nitric 
acid in decomposing silver hyponitrite ; in fact, unless very dilute, it 
is more active in consequence of silver sulphate crystallising out. 
Sodium carbonate (used by Haga and me in 1884, Trans., 46, 78) 
is to be preferred to ammonia for precipitating the salt, as being 
more sensitive, and because the last trace of ammonia is dificult to 
wash out of the silver salt (as Hantzsch and Kaufmann found, 
see p. 114). It is easy to ensure absence of all silver carbonate, 
along with complete precipitation of the hyponitrite, because of the 
solubility of the carbonate in the excess of carbon dioxide always 
present in the solution. 

Digitized by VjOOQIC 


PropertUs qf JSilver HypanUrite. 

Silver hyponitrite is bright yellow, and when pale in colour it 
generally contains a trace of ammonia or loosely combined silver 
oxide. If along with such impurity there is also black silver sub- 
oxide, the colour becomes dull greyish-yellow, but when other 
impurities are absent, the presence of a little black oxide renders it 
somewhat bright green, as seen principally in the crude salt prepared 
by the hydroxyamidosulphonate method. The difference in colour 
observed has even suggested the possibility of the existence of different 
modifications, but there is really nothing to support this notion. 

If precipitated from strongly alkaline solution, or from concen- 
trated solutions of the sodium salt and silver nitrate, or in rubbing 
the calcium salt with strong solution of silver nitrate, silver hypo- 
nitrite is dense, but when precipitated by neutralising its solution in 
dilute acid, it is flocculent and bulky. When deposited from its 
ammoniacal solution through evaporation or large dilution with water, 
it is crystalline (Kirschner ; but also see Faal and Kretschmer). 

It is slightly more soluble in water than silver chloride, and is 
dissolved by very dilute nitric or sulphuric acid, so as to be recover- 
able on quickly neutralising the acid. The nitric acid required to 
dissolve it is considerable, being about 3 equivalents. The sulphuric 
acid solution very soon deposits silver sulphate. Acetic acid 
dissolves it only very slightly in the cold ; phosphoric acid dissolves 
it, but not very freely. It is dissolved by anunonia solution, but only 
sparingly when this is very dilute, and the same salt can be recovered 
either by neutralising or dissipating the ammonia. It is also soluble 
in ammonium carbonate solution, and very slightly in ammonium 
nitrate solution. Qf particular interest is its solubility to a slight 
extent in hyponitrous acid solution, and to a greater degree in 
solution of an alkali hyponitrite. 

It is readily oxidised by strong nitric add. Strong sulphuric aoid 
acts energetically, the heat of reaction being itself quite sufficient to 
decompose some of the salt, a fact which accounts for the production 
of some nitric peroxide and nitrosyl sulphate (see effects of heating, 
p. 108). It is not decomposed by a cold solution of sodium carbonate, 
or by one of Sodium hydroxide if it is weak. It is fully decomposed 
by its equivalent of potassium iodide in solution, but only imperfectly 
by a solution of sodium chloride, if the latter is not in considerable 
excess. When a solution of sodium chloride is shaken with excess of 
undried silver hyponitrite, decomposition ceases when the two sodiuxa 
salts in the solution are in the proportion of 18 eq. of chloride to 
25 eq. of hyponitrite, or, by weight, 4 of chloride to 5 of anhydrous 

Digitized by VjOOQIC 


byponitrite. Absolute alcohol in large excess effects a partial 
separation of the two sodium salts, as already described. 

Silver hyponitrite in the moist state is not entirely stable, for it 
da eo m poeee even at the ordinary temperature, although exceedingly 
slowly ; light and heat qdeken the change, the former modifying it 
to some extent. The decomposition is made evident by the salt losing 
ItB bright oolomr, by its answering to the iodide and starch test for a 
nitrite^ and by its yielding up to water more silver salt (not nitrite, 
but nitrate apparently) than its own very slight solability would 
aecount for. The salt may be washed with boiling water, or even be 
boiled with water, without any very apparent result, but oontinuous 
boiling not only has marked effect in decomposing it, but an action 
which grows in intensity, even though the water is frequently re- 
I^aeed. The water is found to contain silver nitrate, whilst the solid 
salt gives the reaction for nitrite. Masses of moist precipitate retain 
their eoloor outside while drying in a thermostat^ but beM>me greyish 
inside. Silver hyponitrite dissolved in a solution of sodium or potassium 
hypcmitrite decomposes on standing, and very quickly on boiling, 
reduced silver being deposited and sodium nitrate formed in solution. 

Blight diffused light causes enough change in a few hours — bright 
sunlight in a few minutes — ^to allow of nitrite being detected. The 
colour ohange caused by light has been variously described; as a 
matter of fact, in the sufficiently pure salt under water, it is such 
that tlie bright yellow hyponitrite becomes covered with a somewhat 
bright Inown, floceulent substance, very like silver oxide, whicb, per- 
hi^B, it IB ; the blaekening or greying, which has been observed by 
others to be cansed by light, must have been due to impurities, 
although time, as just described, brings about a greying of the salt. 
Silver hyponitrite is least sensitive to light when dry and exposed to 
dry air. Hie main change which occurs in the moist salt, slight as it 
ifly 18 evidently similar to that caused by heat. The salt prepared by 
the hydroxyamidoBulphonate method generally shows an unreal 
stability, due apparently to presence in it of a trace of sulphite, as 
will be explained when the properties of a hyponitrous acid solution 
are treated of; for, in that connection, it has to be taken into con- 
sideration that, like many other precipitated substances, silver 
hyponitrite ia difficult to obtain of high purity. The very slight 
atmoqpherie oxidation of moist silver hyponitrite, described by Haga 
and me (Trans., 1884, 46, p. 78) I now regard as being, not the oxidation 
of the salt itself, but of nitric oxide produced by the very slowly 
deeompoeing salty whieh is then retained as nitrate and nitrite in the 
salt ; the result would be the same as if the wit itself were oxidised. 

Digitized by VjOOQIC 


Effects of Healing Stiver Hyponii/riU. 

In my first paper, it is stated that silver hyponitrite is decomposed 
by a moderate heat into nitric oxide, metallic silyer, and a little silver 
nitrate — ^in this respect resembling silver nitrite, and that it does not 
fuse or exhibit any other change except that from a bright yellow to a 
silver-white colour. That is still a correct statement, so far as it 
goes, but it is imperfect. In 1887» van der Plaats stated that silver 
hyponitrite decomposes explosively when heated ; presumably his 
preparation contained acetate. Thum, who, in 1893, rightly denied 
its explosive character, observed that, in decomposing by heat, the 
bright yellow salt becomes dark brown before assuming the white 
colour of silver, while Elirschner found (1898) that the salt became 
temporarily black. Thum's observation was due, I think, to the very 
dense red, almost opaque, hot nitric peroxide which then poors 
forth, and through which at times the solid mass does look very dark. 
SLirsohner's observation may be due also to this cause, or to his hypo- 
nitrite having contained sulphite. However this may be, the salt 
decomposes with only the change of colour I have described, and in a 
lump of the precipitate the change can be followed by the change in 
position of the sharp boundary line between the bright yellow salt 
and the bright white metal, just as it can be followed in calcium 
oxalate decomposing by heat ; there is no brown or black intermediate 
stage. Thum seems to have found no silver nitrate, but observed the 
production of dense red fumes even when the salt was heated in an 
atmosphere of carbon dioxide, and at a temperature not much 
(f) above 100^. From the important observation of the generation of 
nitric peroxide, he concluded that the decomposition of silver hypo- 
nitrite by heat is probably into silver, nitrogen, and nitric peroxide. 
I had, of course, seen, in my early work, the production of red fumes, 
but had attributed this to the nitric oxide meeting the air, and to the 
decomposition at a higher temperature of the silver nitrate which had 
been formed. The further study of the decomposition which I have 
made has proved that metallic silver, silver nitrate, nitric peroxide, 
nitric oxide, nitrogen, and possibly a trace of nitrite are always produced. 

Having assured myself that nitric peroxide, as well as nitric oxide, 
is evolved by silver hyponitrite when heated, I exposed some to heat 
in a rapid current of carbon dioidde, in order to sweep away as fast 
as I could the nitric peroxide that was produced ; for the production 
of nitric peroidde may sufficiently account for that of silver nitrate 
secondarily. The nitric peroxide and the metallic silver could give 
the nitrate (Divers and Shimidzu, Trans., 1885, 47, 630), but it seems 
improbable that these two substances being produced would then 

Digitized by VjOOQIC 


immediately interact at the same temperature. There is, however, no 
reason why the nitric peroxide of the decomposed part of the salt 
should not act on the undecompoeed portion and thus produce nitrate, 
such interaction readily taking place. My experiment recorded above 
was instituted to see whether I could not almost prevent the forma- 
tion of nitrate. The attempt failed, for I found silver nitrate in the 
residue equivalent to as much as ^th of the total silver, but this 
result does not disprove that the nitrate really is formed in the way 

The nature and composition of the gaseous products were ascertained 
by heating the salt in a vacuum. The quantity of salt taken was in 
each experiment so proportioned to the capacity of the little flask or 
bulb in which it was heated that the volume of the gases at the 
common temperature and pressure should be a little less than the 
capacity of the bulb. The air was removed from the bulb holding 
the salt by means of the mercury pump, while the bulb was kept in 
boiling water to ensure the dryness of the salt. When exhausted, 
the bulb was sealed off, and the silver hyponitrite decomposed by 
heating the bulb in a bath. Thus heated in the absence of air and 
moisture, the salt exhibits scarcely any change below 140°, and only 
slow decomposition between 140° and 160°, but above these tem- 
peratures the change is soon complete. The metallic silver is slightly 
caked together, presumably by the silver nitrate, and the gases are 
faintly red between 140° and 150°, and orange-red at 160° and above. 
On allowing the vessel to cool, the gases become colourless, but regain 
their colour just as before when the vessel is again heated, and these 
changes can be repeated any number of times. 

To examine the contents of the bulb when cold, its point was 
broken off under water, and the small rise of water into the neck of 
the bulb marked ; then the bulb was transferred to a small trough of 
strong solution of sodium sulphite in order to absorb the nitric oxide 
(this vol., p. 82). After an hour or longer, the residual gas was 
examined and measured by bringing the bulb mouth upwards, testing 
the gas as to odour and power to support combustion« and then filling it 
with water from a burette up to the mark already made, and afterwards 
to the mouth in order to learn the volumes of the gases when corrected 
for temperature and pressure. The volumes could be only approxi- 
mately measured in this way, but quite well enough for the purpose. 
1%e metallic silver was weighed, and from its weight and that of the 
hyponitrite, that of the silver nitrate became known. In one experi- 
ment, the bulb was at once freely opened to the air, and the gases 
rapidly blown out ; in this way, the nitric oxide showed its presence 
by reddening in the air, and both the silver and the silver nitrate were 
directly determined. 

VOL. LXXV. Digitized by Google 


These experiments established the production of nitrogen, as well 
as that of the other substances, and the non-production of any appre- 
ciable quantity of nitrous oxide. The quantitative results were that, 
when the decomposition is slowly effected^ as between 140^ and 160^, 
silver hyponitrite yields about 27 per cent, of its nitrogen in the free 
state, and about 20 per cent, when the decomposition is rapidly 
accomplished at higher temperatures. The silver nitrate was formed 
in quantities corresponding with those of the nitrogen, according to 
the equation d(AgON)2»4Ag + 2AgN03 + 2N3, but that, of course, 
proved nothing, since the whole of the nitrate might have been 
formed by the nitric peroxide during the cooling, as certainly much 
of it must have been. On the other hand, the limited quantities of 
nitrogen generated gives full proof that much nitric oxide is either 
primarily formed or comes from interaction between hyponitrite and 
pero3dde, besides what undoubtedly comes from the interaction of the 
nitric peroxide and metallic silver during the cooling. Were none of 
the nitrogen of the salt to become nitric oidde, the free nitrogen 
would be half of the total nitrogen, instead of only three- or four- 
fifteenths as found. 

From the facts observed, it seems to me to be highly probable that 
silver hyponitrite decomposes into silver, nitrogen, and nitric peroxide, 
according to the equation 2(AgON)2e4Ag + N2 + 2NOs, and that 
interaction then occurs between the hyponitrite not yet decomposed 
andsomeof thenitricpero3dde, thus : (AgON)^ + iNO^ « 2AgN03 + 4NO, 
and, therefore, that the decomposition of silver hyponitrite into silver 
and nitric oxide does not occur directly. 

It remains to explain the absorption and regeneration of nitric 
peroxide by cooling and heating the gases in contact with the solid 
residue of the decomposed hyponitrite. The interaction of silver and 
nitric peroxide in the cold, already referred to, explains the disap- 
pearance of the nitric peroxide, half of its nitrogen becoming nitrate 
and half nitric oxide. The regeneration of nitric peroxide at such 
low temperatures as those in the neighbourhood of 150° is explained 
by experiments of mine recorded in a separate note (this vol., p. 83). 
The silver nitrate and nitric oxide interact to produce nitric peroxide, 
andatfirst nitrite, but ultimately silveritBelf,AgNOs + NO » Ag + 2NO2. • 

As to the ExUtmice qf Silver Nitrit<hhyponitrtte, NUrato-hypanUriie^ 
and Ni^aUHnitrUe. 

Silver NitraUHiitriie, — I have made new experiments on the union 
of silver nitrate with silver nitrite, first examined by me in 1871 
(Trans., 24, 85). Silvel' nitrite, mixed with a little less than its 
equivalent of silver nitrate, suffers only slight decomposition until it 

Digitized by VjOOQIC 


melts idong with the nitrate at about 130^. The fused salts solidify 
at about 125° to a translucent, greenish-yellow, crystalline nuu3S, 
except in the uppermost part, where it is opaque from the presence of 
bubbles and metallic silver. This upper part removed, the rest can 
be fused again without suffering further change, and even be heated 
nearly to 180° without decomposing. Silver nitrite, heated alone, 
shows marked change of colour when the temperature has reached 
120% gives red fumes at about 140°, and very freely decomposes 
below 180° without showing signs of fusing. Silver nitrate does not 
fuse below 217° (Carnelly). The low melting point of the mixture of 
the two salts, and the increased stability of the nitrite, are, however, 
the only facte showing that there is any chemical union, for water 
separates the two salts. 

Ifon^asiatence qf NUnUa-hyponiPrite. — Silver hyponitrite (4 parts) and 
silver nitrate (5 parts), in intimate mixture, wei'e heated in a bath. 
Ko change was observed until 175° was reached, when fusion and the 
evolution of red fumes occurred. The hyponitrite had then disap- 
peared, and the fusion may be attributed to the decomposition of the 
hyponitrite, as usual, into nitric oxide, among other things, and to 
the interaction of this nitric oxide with some of the nitrate to form 
the fusible nitrato-nitrite. 

The attempt was also made to prepare a compound of the two salts 
in presence of water, there being some grounds to expect success. 
Galcium hyponitrite, a nearly insoluble salt, was ground up with 
excess of a very concentrated solution of silver nitrate, and a dense 
and strongly yellow precipitate obtained, which was washed with 
water until all the calcium salt had been removed ; the precipitate 
was still yielding up a little silver nitrate when the washing was 
stopped. Drained on a tile and dried in a vacuum, it proved to be 
somewhat sensitive to light and to heat, but, as it. contained 76*94 
per cent, of silver, and could have been washed more free from silver 
nitrate, a combination of the two salts stable in water does not exist. 
All that can be said is that silver hyponitrite requires long washing 
to remove the last portions of silver nitrate. 

NitriUhhyponitrUe ako nanrexiaterU. — In a paper already referred 
to, I have recorded obtaining a minute quantity of what appeared to 
be hyponitrite, when partially decomposing silver nitrite by heat, 
that is, a bright yellow substance insoluble in water and soluble in 
ammonia. I have failed to get this again. Silver hyponitrite and 
sOver nitrite, heated together, show no change until decomposition 
and the escape of red fumes occur, and then all hyponitrite has been 

When making known his observation of the interaction of hydroxyl- 
amine and nitrous acid in 1893, Paal stated that, from a solution of 

Digitized by VjiOOQIC 


alkali byponitrite which also contained nitrite, silver nitrate had pre- 
cipitated a substance which, althongb it was like silver byponitrite, 
proved to be a silver nitrito-byponitrite. It gave no silver nitrite, even 
to hot water, and could be dissolved in cold dilute nitric acid, and be 
reprecipitated with ammonia without suffering change in composition. 
It was less stable than the simple byponitrite when heated, gave the 
reactions of a nitrite along with those of a byponitrite, and yielded 
numbers (not quoted), on analysis for silver, which agreed nearly 
with that required by the formula AgjN^Og. Ten years previously, 
Berthelot and Ogier, probably under similar conditions, got similar re- 
sults, except that they were led by their analysis to give the formula 
Ag^N^Og to the substance they had obtained. It is true that, in spite 
of endeavours to purify it, silver byponitrite retains with obstinacy 
enough nitrite to give the iodide and starch reaction for a nitrite, and 
that it often, through the presence of impurities, gives low results for 
the silver ; but, beyond these admissions, I cannot subscribe to the 
accounts given by the chemists just named as to the existence of com- 
pounds of silver byponitrite with silver nitrite. 

I have reduced sodium nitrite by sodium amalgam as usual, and 
dissolved in the solution one-sixth as much more sodium nitrite as had 
been reduced, thus getting byponitrite and nitrite together in solution 
in about equivalent proportions, in accordance with the experience 
recorded in this paper. The precipitation of silver byponitrite was 
then proceeded with, in one experiment, without previous neutralisa- 
tion of sodium hydroxide, and in another experiment after neutralisa- 
tion of the alkali. The result was the same in both experiments. 
There was a bright yellow precipitate, not noticeably different from 
ordinary hyponitrite, and the mother liquor retained much alkali 
nitrite or silver nitrite in the respective cases ; the precipitate was 
repeatedly washed with cold water, but the washing was stopped when 
very little silver was being extracted. It proved to be somewhat 
sensitive to light and heat. It was dried in the cold and in a vacuum, 
and the silver was then determined ; this was 769 per cent. Nitrite 
could be easily detected in it, but the compound AgjNjO, would have 
only 74 per cent, of silver, and Ag^^fi^ only 76 per cent. Besides 
this, by prolonged washing the byponitrite can be made much purer. 
These experiments, therefore, afford no evidence of the existence of 
such a compound as Paal has described. 

Fropertiea of a Solution qf Hyponitrous Acid. 

Solutions of hyponitrous acid are always prepared in one way, 
namely, by decomposing silver hyponitrite with just sufficient dilute 
hydrochloric acid. Hyponitrous acid has been obtained by Hantzsch 

Digitized by 



and Kaufmann in crTstals very deliquescent and very unstable, by 
using dry ether in place of water in its preparation. The acid in 
dilute aolution reddens litmus not so strongly as nitric acid, but much 
more than carbonic add. On drying the reddened litmus paper, it 
becomes blue again. A solution of the acid becomes neutral to 
litmus when half the quantity of baryta water or alkali required to 
form the normal salt has been added (Zorn), and such a solution, by de- 
composition, soon acquires the 4)roperty of blueing red litmus paper. 
When neutral to litmus, the solution is also neutral to phenol- 
phthalein (Thum). When neutralised with baryta, and very rapidly 
evaporated under reduced pressure, hyponitrous acid yields an acid 
salt which is crystalline and extremely unstable (Zorn). 

It decomposes silver carbonate, if not also lead and other car- 
bonates ; it also decomposes silver nitrate and sulphate. It does not 
oxidise hydrogen iodide (iodide and starch reagent), and is not 
oxidised by iodine solution or by the air. It is oxidised by nitrous 
add and the stronger oxidising agents. No way of deoxidising or 
hydrogenising hyponitrous acid is known ; it entirely resists the 
action of sodium amalgam, and also, according to Thum, that of 
zinc and sulphuric acid. Ethylic hypomtrite is reduced, apparently, by 
tin and acetic or hydrochloric acid to alcohol and nitrogen, according 
to Zorn, but as, also according to him, it slowly decomposes by itself, 
when moist, into nitrogen, alcohol, and aldehyde, there is sufficient 
reason to doubt that this reduction by tin and acid is anything more 
than the hydrogenisation of the aldehyde. 

Hyponitrous acid slowly decomposes into nitrous oxide and water. 
A strong solution soon effervesces, gently in the cold, freely when 
heated, just like a solution of carbon dioxide, and some hyponitrites 
in presence of only a little water effervesce with an acid. A solu- 
tion of one or two grams of the acid in a litre of water kept in ice 
hardly falls noticeably in strength in one hour^ but at 25 — 30° it 
may lose a sixth of the acid by decomposition in 24 hours ; at a 
lower temperature, Thum observed a loss only half as great in the 
same time. Alkali hyponitrites in solution also decompose into 
nitrous oxide and alkali, gradually in the cold and rapidly when 
heated ; alkali hydroxides impede the decomposition, and when highly 
concentrated stop it altogether, apparently (see p. 98) ; neutralisa- 
tion of the alkali even by carbonic acid hastens the decomposition, 
as a matter of course, but there is no evidence that carbonic acid is 
able to decompose a hyponitrite, as it has been said to do. 

Hyponitrous acid solution dissolves silver hyponitrite slightly. 
The alkali salts of hyponitrous acid dissolve silver hyponitrite some- 
what more freely, and also decompose silver chloride (see p. 102) ; 

Digitized by VjOOQIC 


they give precipitates with barium and calcium salts, and with 
solutions of most metallic salts. 

Other substances are liable to be present in the solution of hypo- 
nitrous acid, and this fact has certainly caused the properties of the 
acid to be wrongly described in some respects. In one point, this is 
the case in my first paper, in which, however, there was a warning 
that the crude solution of the acid, which had been examined, might 
have reacted as it did partly through the presence of other unrecog- 
nised substances in it. That solution decolorised iodine water, 
and prevented the action of nitrous acid on an iodide ; this, however, 
was not due to the hyponitrous acid, but to a very little hydroxyl- 
amine, the presence of which was unrecognised. Kirschner has again 
given to hyponitrous acid the property of decolorising iodine water 
to a slight extent. In his case, the substance acting on the iodine 
must have been a trace of sulphur dioxide, for he made his solution 
of the acid from silver hyponitrite that had been prepared by the 
hydroxyamidosulphonate method. I can confirm the accuracy of his 
observation. Even when the silver hyponitrite has been most care- 
fully precipitated so as to avoid all sensible precipitation of sulphite, 
and has been dissolved in dilute acid and reprecipitated, it still 
gives a solution of hyponitrous acid capable of acting on a very little 
iodine water ; but then no more iodine was taken up although there 
was hyponitrous acid in the solution. On the other hand, the 
acid prepared from silver hyponitrite not derived from hydroxyamido- 
sulphonate does not decolorise iodine at all, as Thum first pointed out. 
Hyponitrous acid, according to van der Plaats, liberates iodine 
from potassium iodide ; according to Thum and my first paper, it does 
not, whilst according to Hantzsch and Kaufmann it is only just at 
first that it does not do so. The last-named chemists, therefore, state 
that the acid does not itself liberate iodine, but quickly begins to 
yield nitrous acid which does liberate it. They also found hypo- 
nitrous acid to yield ammonia, but iivalater publication Hantzsch and 
Sauer state that the ammonia was an impurity in the silver hypo- 
nitrite from which the acid had been prepared. Even with the 
simultaneous formation of the ammonia, it is difficult to understand 
the generation of nitrous acid. These authors, invoking the aid of 
tautomerism, suppose that the hydrogen leaves oxygen for nitrogen, 
giving the unknown substance HNIO, which then beoomes NH3 -f 
N2O3, and these again pass into HNOg + N^ + OHj. In place of this 
series of improbable — I would say, unnatural — changes, I suggest 
that, if indeed such change occurs at all, it must be into water, nitric 
oxide, and nitrogen, the nitric oxide then oxidising to nitrous acid. 
But I am strongly disposed to deny that hyponitrous acid decomposes 

Digitized by VjOOQIC 


of itself into anything but what are certainly its main products, 
nitrous oxide and water. My reasons are several. First, there is the 
unlikelihood that the diazo-radidei NIN, should resolve itself into 
mono-nitrogen compounds, such as NO,NH,,NO*OH or {VO)fif in- 
stead of Q^^yy, Secondly, there is the fact that time comes in as the 
condition of the production of nitrous acid, and that a rise in tem- 
perature does not. A solution of hyponitrous acid of fair purity, 
if boiled or quickly evaporated gives nothing but nitrous oidde and 
water ; and only very slowly and to a very small extent does nitrous 
acid appear in a cold solution of the purest acid. Thirdly, the greater 
the care taken to reduce and exclude all nitrite in preparing the 
hyponitrous add solution, the longer will be the time before any 
sensible quantity of nitrous acid develops, and the more gradually 
will the quantity increase. From these facts, the almost necessary 
inference is that the whole of the nitrite has never been entirely 
removed or excluded in preparing the acid, and that what has been 
left, although too minute in quantity to afEect the iodide test (which 
requires 1 in 20 millions, according to Warington), yet multiplies 
itself by interaction with the hyponitrous acid, forming nitric oxide, 
which is further oxidised to nitrous acid by the air dissolved in the 
solution (HNO)j + 2HNO, = 2H20 + 4NO-*4HN02. This aerial oxi- 
dation can be demonstrated in such a solution of hyponitrous acid as 
that which Hantzsch and Kaufmann employed in their experiments, 
which gave the blue of the iodide test almost immediately ; it is only 
necessary to leave one portion of the solution in a deep, narrow vessel, 
such as a test tube half full, and another portion in a shallow basin for 
10 minutes, and then apply the test, when the solution in the basin will 
be found to liberate more iodine than that in the tube. If, in reducing 
the sodium nitrite, its concentrated solution is shaken with excess of 
the amalgam for an hour or two after its main reduction, and the 
solution is then either diluted, acidified cautiously with dilute sul- 
phuric add, and tested, or is predpitated by silver sulphate, away 
from the light as far as practicable, and the predpitate washed in the 
dark and converted into the add and tested, either solution, when 
mixed with the iodide reagent, will not blue in the least for an hour 
or more in the dark, and provided the constituents of the reagent are 
pure enough and properly used.* 

Against the view, which may be advanced, that hyponitrous acid 
becomes nitrous add through oxidation by the air, I must point out 

* My way of applying the test is that followed by Warington (Ohenu Nitws, 
1885, 61, 89), except that, having potassimn iodide of high quality, I need it 
instead of l^mmsdorfs zino iodide solution. In the dark, a blank test will remain 
for houB without the least blueing. There is no advantage in uiing acetic acid 
in plaee of pnn sulphurio or hydrochloric add. 

Digitized by VjOOQIC 


that it is difficult to admit that if the nitrous acid has such origiui it 
should form so very slowly. A way occurred to me which must be used 
for deciding this point, so far as the exclusion of nitrous acid goes, but 
it has, in my opinion, not served to do so. If, in preparing sodium 
ozimidosulphonate, the sulphur dioxide is used in excess, every trace 
of nitrite ought, presumably, to be sulphonated ; if, then, the oximido- 
sulphonate is fuUy hydrolysed into hydroxyamidosulphonate, as it 
presumably can be, then, when the latter is converted into hyponitrite 
and sulphite by potassium hydroxide, there will be no oximido- 
sulphonate present to simultaneously revert to nitrite and sulphite. 
Therefore, the silver hyponitrite from such a source should be obtain- 
able absolutely free from nitrite, and should furnish a solution of 
hyponitrous acid also free from nitrous acid. Such silver hyponitrite 
I endeavoured to prepare, and then tested the acid got from it. The 
issue was, however, complicated by the fact that such an acid is not 
quite free from sulphurous acid, as was shown by its decolorising a 
minute quantity of iodine solution. That it also did not act for a time 
on the iodide and starch reagent was due in part to this cause. The 
solution did, however, give a blue coloration with the reagent sooner 
than a corresponding blank test. But this was no proof that hypo- 
nitrous acid passes spontaneously into nitrous acid, for, first, there is 
the possibility of nitrous acid having been present through incomplete 
sulphonation and hydrolysis in preparing the hyponitrite ; this nitrous 
acid would, indeed, have been converted into nitric oxide by the 
sulphurous acid retained by the silver salt, but when all this was 
gone, the nitric oxide would have become nitrous acid again by 
oxidation. Secondly, it is almost certain that the oxidation of the 
sulphurous acid by the air would have induced oxidation of some hypo- 
nitrous acid, in accordance with the observations of Mohr, M. Traube, 
van't HofE and Jorissen, Engler and Wild, Bach, ^. 

QuaTUitative Estimation qf ffj^ponitrous Add. — Hyponitrous add can 
be estimated accurately, both gravimetrically (Zom) and volumetrically 
(Thum). Solutions of the free acid, or of its alkali salts in water, or 
of its other salts in very dilute and cold nitric acid are mixed with 
excess of silver nitrate, and then the free acid is just neutralised with 
sodium carbonate or with ammonia. The washed precipitate is either 
dried and weighed as such, or weighed as metal or as chloride. 

Yolumetrically, the acid can be estimated in solution in the free 
state and unmixed with any other acid, by adding excess of solution 
of potassium permanganate, leaving it for a quarter of an hour, then 
adding sulphuric acid, allowing it to remain for another quarter of an 
hour, warming to 30^, adding a known quantity of oxalic acid sufficient 
to decolorise, and, finally, titrating back with permanganate. The 
hyponitrous acid is thus oxidised to nitric acid. The oxalic acid should 

Digitized by 



be decinormal, and the solution of permanganate be volumetrically 
equivalent to it. Ferrous sulphate is unsuitable for use in place of 
oza&ic acid. The process is an excellent one. Hantzsch and Sauer 
failed to get good results, because they deviated from Thum's diiec- 
tiona by acidifying the permanganate before adding it to the hypo- 
nitrite. Elirschner also was unsuccessful with this process, but his 
failure is also explained by his deviation from Thum's directions. He 
added nearly insoluble salts, such as the barium, strontium, or silver 
hyponitrite, to the potassium permanganate, so that the base of the 
salt was present, and the hyponitrous acid locally in excess of the per- 
manganate ; he then added the sulphuric acid, apparently immediately! 
and used ferrous snlphate for titrating back. 

Taking 5 c.c. of normal hydrochloric acid, largely diluting it, adding 
ice and a cream of precipitated silver hyponitrite so as to exactly use 
up all the hydrochloric acid, making up to 100 c.c, and 'decanting 
from the bulk of the silver chloride, I obtained a solution which, 
although somewhat turbid from silver chloride, gave, in successive 
portions of 20 cc, all tested within an hour, quantitative results 
corresponding well with 0*155 gram hyponitrous acid in 100 c.c, that 
isy the quantity equivalent to the hydrochloric acid taken ; next day, 
the remainder of the solution (in very hot weather) showed the presence 
of 0-131 gram of the add in 100 cc. 

Thum found that, in alkaline solution, alkali hyponitrite is quanti- 
tatively converted into nitrite by permanganate. Although I have 
not examined this point myself, I find that nitrite is thus formed, 
and that nitric acid is formed in Thum's acid permanganate method. 
Kirsehner doubts that either is produced. 

Barium, Stnmtiwn^ and Calcium HyponUritea. 

Barium hypofUkite has been obtained by Zorn, Maquenne, and 
Kirsehner, and is most simply prepared by adding barium chloride to 
a concentrated solution of sodium hyponitrite and stirring well. It 
is crystalline, almost insoluble, and an unstable and exceedingly 
efflorescent salt, but Kirsehner has succeeded in determining its water 
of eiystaUisation satisfactorily. Its formula is BaNgOj + iHjO. A 
crystalline acid salt exists (Zorn). 

Straniiufn hyponitrite^ SrNjOj + SH^O, Maquenne, Elirschner. 

Caleivm hyponitrite, CaNjOj + iH^O, Maquenne, Kirsehner.* This 
is crystalline, very sparingly soluble, stable, not losing its water even 
over sulphuric acid. I find that it can be easily precipitated from a 
fairly concentrated solution of sodium hyponitrite, and can thus be 
prepared more easily than in the ways followed by Maquenne and by 

Digitized by VjOOQIC 


Eirschner, using the silver salt. On account of its stabilityi it is a 
good hjponitrite to keep in stock. It is sufficiently soluble for its 
solution to serve to show the reactions of a hyponitrite with silver, 
mercuric, mercurous, copper, lead, and other salts. 

Calcium, Strontium, and Barium HyponUrosoacetaUi. 

Some remarkable salts have been described by Maquenne, having the 
composition expressed by the f ormulsa 

BaN,O^Ba(C,H302)2,(C,H,0,)2+ 3H,0. 
T have prepared and partly analysed the calcium salt, following 
Maquenne's process, which is to dissolve calcium hyponitrite in 30 per 
cent, acetic acid until the new salt begins to crystallise out. I kept 
the acid at about 50° while dissolving in it nearly as much of the 
calcium salt as it would take up, the salt being deposited on cooling. 
It is remarkable that this can be done without causing more than very 
slight effervescence. The salt crystallises in short prisms, stable for 
many days, very soluble in water, in which it gives, with silver nitrate, 
the yellow hyponitrite. In spite of its acid composition, it is neutral 
or even slightly alkaline to litmus. To account for its existence and 
neutral reaction, I suggest for it the constitution expressed by the 

Ca{OAc)j.Ca<g^^«' ^^° + 5H,0. 

This represents it as being normal calcium acetate with one-fourth oi 
its oxylic oxygen replaced by the hyponitrite radicle, or as a doubl 
anhydride of calcium acetate and hyponitrite. It is thus made- out t 
have a constitution analogous to that of a (hypo)nitroso8ulpbate, a 
determined by Haga and me (Trans., 1895, 67, 1098), 

Simple hyponitrosoacetic acid, C^HgO'O 'N^* OH, would be isomei 
with acetonitrosohydroxamic acid,'^C2H30-N(NO)*OH, which Santzs 
and Sauer have been trying to prepare. The hyponitrosoaoetal 
are much more stable in water than the (hypo)nitrososulpliate8, 
difference perhaps connected with the fact that sulphuric acid ioni 
largely, while both acetic acid and hyponitrous acid ionise very litl 
Heated with water, the hyponitrosoacetates decompose like tlie (by] 
nitrososulphates do in cold water. 

Digitized by VjOOQIC 


Mercuric Hypanitrite, 

Mercuric hyponitrite is a particularly interesting salt and has^not 
as yet been described. R&y has, indeed, described some compounds 
vbicb be regards as being basic mercuric hyponitrites, but obtained 
under conditions suggesting the probability that they are something 
quite different ; moreover, he has not as yet proved them to be com- 
pounds of this class. One of them he obtained by the interaction of 
solutions of mercuric nitrite and potassium cyanide, a very interesting 
and remarkable result, should it be confirmed. In any case, his pre- 
cipitates appear to have nothing at all in common with the normal 
salt here described, and cannot be obtained in thd ordinary way. The 
existence of this salt was indicated by me in 1871. 

Mercuric hjrponiirite is obtained from sodium hyponitrite and 
mercuric nitrate by precipitation ; the solution of sodium hyponitrite 
and hydroxide, obtained by reducing sodium nitrite, is largely diluted 
and, while ice cold, nearly or quite neutralised with dilute nitric acid ; 
it is then (mercurous nitrate serving as indicator, see p. 97) poured 
into a mercuric nitrate solution, which must not be in excess and 
should contain as little free acid as possible. The slightly turbid 
mother liquor is quickly decanted from the precipitate formed, and 
after being neutralised with sodium carbonate is mixed with more 
mercuric nitrate, the whole poured back on to the main precipitate, 
stirred up with it, and soon again decanted. The precipitate should 
be washed quickly by decantation, since it is liable to be quickly 
destroyed by the slightly acid mother liquor. 

It is a flooculent, cream-coloured precipitate, easily washed on the 
filter, and dries up to a light buff-coloured powder, this colour being 
due, probably, to incipient change into the mercurous salt. Dried 
quickly in the air, on a porous tile, it is hydrated, having the formula 
{Qjg^fi^^'\-Z'S.fiy but if dried in the desiccator it is anhydrous. 
Being a little sensitive to light, it should be dried in the dark. It 
dissolves in hydrochloric acid and in sodium chloride solution, but it is 
unstable, changing into the mercurous salt, and, therefore, is liable to 
show turbidity in the chloride solutions. The mercury, precipitated as 
sulphide from a solution of the anhydrous salt in hydrochloric acid, 
was found to be 76*71 per cent., the formula HgNgOj requiring 76*92 
per cent. Its solubility in excess of sodium chloride does not prevent 
mercuric chloride giving a precipitate with sodium hyponitrite. The 
solubility of the salt in sodium chloride is a qualitative proof of its 
normal composition. The alkalinity of the solution is caused by the 
sodium hyponitrite generated in it. Potassium hydroxide at once 
decomposes mercuric hyponitrite into oxide, without showing any 

Digitized by VjOOQIC 


tendency to produce basic salts. In very dilute alkali, the precipitate 
is slightly soluble. 

What makes this salt so remarkable, not only as a hyponitrite, but 
as a mercuric salt, is the nature of the decomposition which it under- 
goes. Slowly or quickly, it decomposes into mercurous hyponitrite 
and nitric oxide — some of the latter, oxidised by the air, converting 
some hyponitrite into nitrate. No other mercuric salt decomposes 
into mercurous salt, although many cupric salts change into cuprous 
salts. Ferric oxalate shows just the same kind of change, namely, 
into ferrous oxalate and carbon dioxide. The most closely related 
change, however, is that of sodium (hypo)nitrososulphate into sulphite 
and nitric oxide, the very phenomena being similar, so that, except for 
the colour change, I might describe my experience with this salt in the 
words of the paper by Haga and me on sodium (hypo)nitrososulphate 
(Trans. 1895, 67, 1095). Thus, having on one occasion left some 
grams of salt all night in the desiccator in the form of a pressed cake, 
just as removed from the filter, I noticed, when weighing it between 
watch-glasses, that it was losing weight on the balance-pan. When the 
glasses were opened, a strong nitrous odour was observed, the cake soon 
became grey, white on the surface, and, being left loosely covered, grew 
very hot and gave out torrents of nitric oxide ; it then cooled, and 
underwent no further change, even in the course of months. The 
whitish colour of the cake was found not to penetrate beyond a milli- 
metre into it, the inside being of a uniform yolk-yellow, and consisting 
of mercurous hyponitrite. The surface-coating proved to be mercurous 
nitrate, largely soluble in water, and had evidently been produced 
with the assistance of the oxygen of the air. Not always, however, 
does the change occur in this striking and rapid way, its progress 
being gradual and almost imperceptible until complete. 

Mercuric hyponitrite is decomposed by heat largely into mercuric 
oxide and nitrous oxide, but partly into metal and nitric oxide. 

Other Hyponitriiea, 

Mercwroua Hyponitrite, — This salt has been prepared and analysed 
by Thum (Inaug, Bits., Prag. 1893), who used sodium hyponitrite and 
mercurous nitrate in obtaining it. The possibility of getting it by the 
spontaneous decomposition of mercuric hyponitrite has just been 
described. RAj has also evidently obtained it in a very impure state, 
not further examined. It can be prepared in the same way as 
mercuric hyponitrite, using mercurous nitrate in place of mercuric 
nitrate. It is of a full yellow colour, is blackened by even the 
weakest solution of alkali, and is soluble in dilute nitric acid, from 

Digitized by VjOOQIC 


which it can be precipitated by sodium carbonate. It is a stable salt, 
but is blackened by bright light. Its decomposition by heat resembles 
that of the mercuric salt, except that much more metal is produced, 
as is natural Composition, (ngON)2. 

Cupric Hydroxide Eyponitrite. — ^This salt was described by me in 
1871, and was also obtained by Kolotow in 1890, but was first fully 
examined by Thum, and has again been examined by Kirschner; 
being a basic salt, its precipitation from normal sodium hyponitrite 
leayes an acid mother liquor, on neutralising which much more of the ' 
salt precipitates. It is of a bright pea-green colour, and very stable. 
It may be boiled with water without losing its colour, but is decom- 
posed by sodium hydroxide and is soluble in dilute acids and ammonia. 
Thum has shown its composition to be Cu(OH)NO. It gives water, 
cupric and cuprous oxides, and nitrous and nitric oxides when heated. 
By adding copper sulphate in excess to hydroxylamine sulphate and 
then a very little ammonia, it can also be precipitated in small 

Cuprous hyponiirUe cannot be formed. I have tried to get it by 
precipitating sodium hyponitrite by copper sulphate in presence of 
free hydroxylamine, but, first, cuprous oxide precipitated and then, 
by atrial oxidation, the basic cupric hyponitrite, which in composition 
is equivalent to that of cuprous hyponitrite combined with faiydroxyl 
(see above). 

Lead ffyponiiriie. — ^This salt was also briefly described by me and 
has been prepared and analysed by Thum ; Kirschner has again pre- 
pared and analysed it, but not in a pure state. The precipitate is 
cream-yellow and flocculent, but soon becomes very dense and sulphur- 
yellow ; its first state is probably that of a hydrate ; Kirschner has 
mistaken it for a basic salt. The yellow salt is PbNgO,. As Thum 
has pointed out, the yellow precipitate, when formed in a weak acid 
solution, is crystalline and just like ammonium phosphomolybdate. It 
is soluble in dilute nitric acid, and is decomposed by sodium hydroxide, 
but not by sodium carbonate in the cold. 

Ammonium Hydrogen Hyponitrite. — ^This salt has been described by 
Hantzsch and Kaufmann, who found it to be exceedingly unstable, as 
was to be expected. That the normal salt could not exist had already 
been pointed out by me, and by Zorn ; D. H. Jackson believes, how- 
ever, that he did obtain it in small quantity in prismatic crystals, but 
this is exceedingly improbable. 

Ethylic Hyponitrite. — ^This alkylic salt was prepared by Zorn, and its 
vapour-density taken by him. It is very explosive, and is not saponi- 
fied by potassium hydroxide. In the moist state, it slowly decomposes 
into nitrogen, alcohol, and aldehyde. 

Benzylie Hyponitrite. — Hantzsch and Kaufmann have prepared 

„.gitized by Google 


benzylic hyponitrite and determined its molecular magnitude cryoscopi- 
cally. It undergoes similar decomposition to the ethylic salt. 

Constitution of the ffypomiritea. 

MoleciUar Magnitude, — In my first paper, nothing could be said as to 
the molecular magnitude and constitution of the hyponitrites. In 1878, 
Zorn fully determined their molecular magnitude, finding it to be that 
containing 1^^* ^^^^y by establishing the existence of an aoid barium 
salt and illustrating the similarity of hyponitrites to carbonates 
(a point which had already been noticed by me), and then 
by preparing ethylic hyponitrite and taking its vapour-density at 
reduced pressure (Hofmann^s method). It would, therefore, be unjust to 
the memory of this x^hemist to admit Hantzsch's claim {Annalen, 1898, 
299, 68) to have finally established this point by determining cryoscopi- 
cally, in conjunction with Kaufmann, the molecular magnitude of 
hyponitrous acid in water and of benzylic hyponitrite in acetic acid, 
valuable as these determinations are. The possibility of determining 
the molecule of the acid in its solution in water rests upon the fact, 
also ascertained by these chemists, that the acid only slightly ionises 
even in very dilute solution. Maquenne, by a somewhat uncertain 
form of the cryoscopic method, has also shown that, in calcium hypo- 
nitrosoacetate, the hyponitrite radicle cannot be less than N^Oj. The 
strong alkalinity of the alkali salts, and the want of action on litmus 
of their partially-neutralised solution, first pointed out by me, and 
the solubility, although only slight, of silver hyponitrite in hypo- 
nitrous acid solution (Thum) and in alkali hyponitrite solution, are 
also facts in accordance with the dihydric composition of the acid. 
Other chemical evidence of the diazo-grouping in hyponitrites is 
afforded by the fact of the difficulty, if not impossibility, of deoxidising 
or hydrogenising them (seep. 113). The derivation of hyponitrites 
from the interaction of hydroxylamine and nitrous acid would only 
afford evidence of the diazo-magnitude of the molecule, if the hypo- 
nitrite produced were much larger in quantity than what can be 
obtained from hydroxylamine by other oxidising agents, or from nitrous 
acid by other reducing agents. 

My colleague. Assistant Professor Ikeda, has kindly made some 
determinations of the molecular magnitude of sodium hyponitrite by 
Loewenherz's method (Zeit. physikeU, Chem, 1896, 18, 70), in which the 
lowering of the freezing point of melted hydrated sodium sulphate by 
another sodium salt is observed ; Loewenherz found that sodium salts 
behave towards the water of hydrated sodium sulphate almost as non- 
electrolytes. Prof. Ikeda, in his experiments, employed sodium 
thiosulphate in place of sulphate, but only because he had been working 

Digitized by VjOOQIC 


with that salty and had had large experience with it. Unfortunately, 
the anhydrous sodium hyponitrite I could furnish at the time was 
contaminated with 4 or 5 per cent, of carbonate (same mol. wt.), so 
that the determination of the molecular magnitude can only be regarded 
as approximate. But it is amply sufficient to decide between NaON = 
53y and (NaON),^ 106, if that were any longer necessary, after Zorn's 
decisive researches, supplemented by those of Hantzsch and Kauf mann. 
Prof. Ikeda has given me the following details. 

M. p. of NajSj03 + 5H30 = 48-4° (Tilden found 48-6'^) ; 

H. of fusion = 42-8 Cal. (Ikeda) ; 

Wt. of thiosulphate used » 40*9 grams ; 

Wt. of sodium hyponitrite used»^ grams ; 

Dp. of solidifying pt. = AT° ; 

MoL wt. of hyponitrite » m. 



2(273 + 48-4)« 

















In cases where no decomposition of the salt occurs, the method gives 
results too high, as, for example, 78*6 instead of 69, for sodium 
nitrite ; but taking into consideration the partial hydrolysis of the 
hyponitrite that certainly takes place, its molecular weight is clearly 
indicated as 106 rather than 53. 

CcngiiHUion, — The constitution (HNO), seems to be excluded by 
considerations of valency, but the positive evidence for (HON)^ is 
ample. Zorn's observation that ethylic hyponitrite decomposes into 
nitrogen and alcohol (and aldehyde), even in presence of reducing 
agents, establishes the diazo-grouping in hyponitrites. Ammonia or 
other amine ^is never produced in the decomposition of any hypo- 
nitrite. Then the conversion of a hydroxyamidosulphonate into hypo- 
nitrite affords a beautiful demonstration of the oxylic constitution of 
the hyponitrites, 

2HO-NH-S08Na+ 4NaOH « NaONINONa + 2NaS03Na + 4H2O. 
Hiantzsch and Sauer have also given an equally convincing proof of 
the same pointy by introducing nitrosyl into dimethylhydroxycarb- 
amide and decomposing the product by alkali (see below). The facts 
that cuprous hyponitrite cannot exist, and, on the other hand, that 
the mercurous hyponitrite, and not the mercuric salt, is stable, point 
also to the metals being united to the oxygen, and not to the nitrogen. 

Digitized by VjOOQIC 


HantzBch and Sauer, in their deBire to prove that nitramine is not 
H^N'NOg, bat a stereoisomeride of hyponitrous acid, would have it 
that their interesting formation of hyponitrous acid from dimethyl- 
hydrozynitroBOcarbamide is analogous to that of nitramine from 

EtO-OO(NjOjH) + HOH = H(Nj,02H) + EtO-OOOH ; 
lOf Oj- C0(N,03H) + HOH « H(N,02H) + NMoj- CXX)H (decomposing). 
However, by displaying what (NjO^H) conceals, namely, the difference 
between the nitramine and the isonitramine, 

Thiele. Hantztch. 

EtO-00 -NH-NOjj, or EtO-00 'N;— ;N- OH + H^O - NH^-NO^, or 

HN— N-OH + Ac. 

NMoj- CO-N(OH)NO + HjO = N(0H):N0H + Ac., 

it becomes evident that the hydrogen of the water (or metal of the 
alkali) goes, in the case of the nitramine, to the amidio nitrogen united 
to the carbonyl, whilst, in the case of the isonitramine, it goes to the 
nitrozy- or nitroso-nitrogen not united to the carbonyl, even if Hantzsch 
and Sauer's free resort to tautomery could be justified. Surely, this 
difference is too great to allow of nitramine being treated as a probable 
or actual stereoisomeride of hyponitrous acid. Hantzsch's formula, 
KO 'Nr^^N' SOgK, for potassium (hypo)nitrososulphate has been shown 

by Haga and me to have nothing favouring its preference to that of 
KO'NIN'O'SOgE, which has so much to be said' for it. 

Analogy of HyponUrites to Carhonatea — of Nj to CO. — In hardly 
forming salts with the feebler metal radicles, such as aluminium and 
f erricum j in decomposing readily into anhydride and water ; and in 
having its soluble normal salts with very alkaline reaction, hypo- 
nitrous acid resembles carbonic acid, as was indicated in my first 
paper. Zom, also, in one of his papers, dwells on the analogy of the 
one acid to the other, pointing out that the salts have the same 
moleoolar magnitude, since Nj and 00 are both 28. 

As is well known, the physical properties of nitrogen and carbon 
monoxides are throughout almost identical. The radicles, carbonyl 
and dinitrogen, also are both bivalent, and occur combined with ozylic, 
imidic, and alkyl radicles. Thus, C0(0Na)2 and COONa(OH) find 
their analogues in N2(ONa)2 and N20Na(0H). Just as ferric oxalate, 
^«2(^2^2^2)s» becomes Fe2(C20302)2 + 2000, so ^g2Q^P%)% becomes 
Hg2(N202) + 2NO. COO corresponds with N^O ; also COINAg to 
NjINAg. Lastly, ketonic compounds are perhaps represented by 

Digitized by VjOOQIC 



H. Davy, Researches, 1800, 254 ; Hess, Ann, Phys. Chem., 1828, 12, 
257 3 Pflouze, Ann. Chim. Phys., 1836, [ii] 60, 151 ; SchoenbieD, 
Cham., 1861, 84, 202; Be Wilde, Butt. Ae. Belg., 1863, [ii], 16, 560, 
and Awn., 1864, SuppL, 3, 175 ; Fremj, dmpt. rend., 1870, 70, 66 and 
1208 ; Maamen6, Compi. rend., 1870, 70, 149 ; /. Ch. Soc., 1872, 26, 
772; Ch. Nms, 1872, 26, 153 and 285; Th^<yrie genh-ale de Vacti<m 
eUmiqus (Paris: Danod), 1880,^286; Divers, Proe. Ray. Soe., 1871, 
10, 425 and, in part, Ch. News, 23, 206 ; Ber., 18^6, 29, 2324 ; Ann., 
1897, 296, 366 ; Divers and Haga, J. Ch. Soc, 1884, 46, 78 ; 1885, 
47, 203 and 361 ; Proc. Ch. Soc., 1887, 3, 119 ; J. Ch. Soc., 1889, 66, 
760 ; 1896, 60, 1610 ; Zorn, Ber., 1877, 10, 1306 ; 1878, 11, 1630 and 
3317 ; 1879, 12, 1509 ; 1882, 16, 1007 and 1258 ; van der Pl^ttsi 
Ber., 1877, 10, 1507 ; Menke, J. Ch. Soc., 1878, 33, 401 ; Berthelot 
and Ogier, Compt. rend., 1883,96, 30 and 84 ; Bertbelot, Compt. rend., 
1889, 106, 1286 ; Danstan and Dymond, J. Ch. Soc., 1887, 61, 646 ; 
Dunatan, Proe. Ch. Sac., 1887, 3, 121 ; Maquenne, Compt. rend., 1889, 
108, 1303; Kolotow, J. ph. Russ., 1890, 23, 3; Abstr. in Chem. 
Centr., 1891, i, 1859, and BuU. Soc. Chim., 1891, [iii], 6, 924 ; Thorn, 
Inaug. Diss., Prog., 1893 and, in part, Monaish., 1893, 14, 294; W. 
Widioenns, Ber., 1893, 26, 771 ; Paal, Ber., 1893, 26, 1026 ; D. H. 
Jackson, Proe. Ch. Soc, 1893, 9, 210 ; Tanatar, J. Russ. Ch. Soc, 
1893, [i], 26, 342; Ber., Ref., 763; Ber., 1894, 27, 187; 1896, 
29, 1039 ; Hantzsch, Ber., 1896, 29, 1394 ; Hantzsch and Kaufmann, 
Ann., 1896, 292, 317; Hantzsch and Sauer, Ber., 1898, 299, 67; 
Pfloty, Ber., 1896, 29, 1659; BAy, J. Ch. Soc, 1897, 71, 347, 1097, 
and 1105; Kirschner, Zeit. anorg. Chem., 1898, 16, 424. 

XVI. — Derivatives of Camphoric Acid. Part II L 

By Febderio Stanley Kipping, Ph.D., D.Sc., F.R.S. 

The first of these papers on derivatives of camphoric acid (Trans., 
1896, 69, 913) contained an account of a number of optically active 
adds and other compounds which had been obtained from ir-bromo- 
camfdioric add, in the course of attempts to open the closed chain 
contained in this simple substitution product of camphoric acid at a 
different point from that at which it is broken in the oxidation of 
euDphoric add, the parent substance, to camphoronic acid. 

The prindpal acids, which have already been described, were 
obtained in the following manner. if-Bromocamphoric acid was treated 
with an aUkali and thus converted into (a) tranfl^nssamphanio add, 

VOL. LXXV. C^r\(^ci\o 

Digitized by VjO(J*vIVL 


Cjo^i4^4> a monocarbozylio acid containing a lactone ring, and 
(5) ir-hydrozycamphoric acid, Cj^H^gOg ; these two compounds wero 
oxidised by nitric acid to one and the same ^aiM-«M3amphotricarbozylic 
acid, from which the oa-isomeride was also obtained. 

The farther oxidation of ^nsnf-Tr-camphotricarboxylic add being 
attended with considerable difficulty, owing to the great stability of 
the compound, it was converted into its monobromo-substitution pro- 
duct and the latter was then decomposed with water ; in this way, two 
isomeric lactones of the composition OioHi^^^, derived from a 
hydroxycamphotricarboxylic acid, O^QK^fi^, were obtained, but these 
derivatives, like the parent substance, were found to be very stable, 
and could not be made to yield satisfactory oxidation products. 

Having failed to obtain any easily oxidisable derivative of cam- 
phoric acid by these means, attempts were next made to oxidise cU-^r- 
camphanic acid, CjqHj^O^, an interesting compound obtained by the 
distillation of its isomeride ^an«^-camphanic acid ; these experiments 
were successful in a measure, inasmuch as a crystalline hydroxy-ctfir- 
camphanic acid was thus obtained, but this compound, like the others 
just mentioned, effectually resisted further oxidation. 

It will be seen from this brief statement that the main object of 
this investigation had been defeated by the stability of these different 
substitution products of camphoric acid, all of which still contained the 
closed chain of the parent substance ; moreover, the difficulty of 
obtaining material sufficient for oxidation experiments on the large 
scale rendered it almost impossible to proceed further on these lines. 
It was necessary, therefore, to try and attack the ir-bromocamphoric 
acid in a somewhat different manner, namely, to convert it, if 
possible, into a dibromo-compound, and then to substitute hydroxyl 
for one, or both, of the bromine atoms, and thus obtain in a more direct 
manner a substitution product which might be oxidisable. 

The results of these experiments are described in this paper. 

When ir-bromocamphoric acid is brominated, it is converted, in 
almost theoretical quantities, into a substance of the composition 
CioHi2^^2^s> which, for reasons given later, is named vw *-dibromocam- 
phoric anhydride ; this substance crystallises in magnificent rhombic 
plates, and on hydrolysis with concentrated nitric acid it is trans- 
formed into nw-dUn'omocamphoric acid, C^QH^^Br^O^, a substance which 
in many respects is very similar to ir-bromocamphoric acid and other 
halogen derivatives of this compound which have been described 
(compare Lapworth and Kipping, Trans., 1897,71, 15). 

nt^Dibromocamphoric acid is decomposed when it is heated alone or 
with water, yielding hydrogen bromide and an acid of the composition 

* The letter w has been previonaly med to denote a certain pocition in the 
molecale of camphoric acid (compare Trana., 1898 69i 61, and 916). 

Digitized by VjOOQIC 


C|^j3Br04, which, from its method of formation, must be regarded as a 
lactonic monooarbozylie add ; for this and other reasons, this compound 
is named w^^iromO'VHXifnphanic add. It crystallises in orthorhombic 
prisms, and both its amid€^ CgHi^BrO^'OO'NH,, and msihylio salt, 
OgH^^rO^-OOOMe, are well-defined compounds. 

When x>bromo-tCHiamphanic acid is treated with alkalis, it yields a 
crystalline product which is identical with hydroxy-cie-n^camphanic 
acid, the oxidation product of cta-vcamphanic acid ; this hydroxy-acid 
is also formed by the action of excess of alkali on nto-dibromocam- 
phoric anhydride. When, on the other hand, ^r^bromo-u^camphanic 
acid is boiled with nitric acid and silver nitrate, it is converted into a 
compound of the composition O^qH^jO^ which is identical with one of 
the iiM>meric lactones of hydroxycamphotricarboxylic acid. 

These transformations, the chemical properties of the various com- 
pounds, and the relations existing between these and other substitu- 
tion products of camphoric acid, can only be elucidated by making use 
of some structural formula for camphoric acid, and in previous papers 
Bredt's formula was employed because it accounted for all the facts 
ondmr discussion in a satisfactory manner, and at the same time 
seemed to have more evidence in its favour than any other of the 
many formuln which had then been suggested. The recent publica- 
tion of an important paper by W. H. Perkin, junr. (Trans., 1898, 73, 
796), in which a new formula for camphoric acid is propounded, ha» 
certainly necessitated a modification of this last statement ; never- 
theless, the difference between Bredt's and Parkin's formulee is of such 
a kind as to require little, if any, alteration in the views which have 
been advanced in explaining the relationship of these ir-derivatives. 

On inspection, it will be seen that these two f ormulsB, which are 
given below, differ merely in this; namely, that one of the >'CH.j 
groups in Bredt's is transferred, in Perkin's formula, to a position 
between the >OMej and >OH*OOOH groups. 



Bredt Perkin. 

This change does not effect any of the author's previous arguments, 
which were mainly based on stereochemical considerations, and, iti 
£act» only makes this difference, that the ir^amphanic acids, according 
to Perkin's formula, may, but do not necessarily, contain a 8- instead 
of a y-lactone ring. 

K 2 

Digitized by VjOOQIC 


Making use, then, of Perkin's formula, and representing v-bromo- 
camphoric acid by I (below), the substitution of bromine for hydrogen 
would afford a dibromocamphoric acid which must be represented by 
formula II, because in the formation of this substance the bromine 
doubtless displaces the same tertiary hydrogen atom as that which is 
expelled in the formation of to-bromocamphoric anhydride from 
ordinary camphoric acid. 

CH,— C<CB^^ CH,— CK^^H 

L r-Bromocamphoric acid* 11. ino-Dibromocamphoric acid. 

ir^Bromo-tiMiamphanic acid, which is formed from the dibromo-acid 
with the greatest readiness, just as to-camphanio acid is very easily 
produced from to-bromocamphoric acid, might be represented by one of 
several isomeric formuln, of which, however, it is unnecessary to give 
examples, as it is possible to select the most probable one from a con- 
sideration of the following facts and arguments. In the first place, 
there are strong grounds for supposing that the v^bromine atom, 
namely, that which is a constituent of the -OHjBr group, is still 
present in this bromocamphanic acid, because it is known that the 
to-bromine atom in lo-bromocamphoric acid is very easily eliminated 
as hydrogen bromide by the action of boiling water, whereas «^bromo- 
camphoric acid is comparatively stable under these conditions ; in the 
second place, the bromocamphanic acid in question offers considerable 
resistance to the attack of oxidising agents, whereas had it been 
formed by the elimination of the n^bromine atom, it should resemble 
fratMircamphanic acid in behaviour and be easily oxidisable to a 
bromo-tricarboxylic acid. 

A comparatively simple way of settling this question as to which 
of the two halogen atoms is eliminated from the dibromo-acid offered 
itself, namely, to prepare a ir-bromo-to-chlorocamphoric anhydride by 
chlorinating the ir-bromo-acid, and then to convert this substance into 
the substituted camphanic acid by treatment with boiling water. On 
making these experiments, it was found that the Tr-bromo-uhehlarO' 
camphoric anhydridcy C^oHj^BrClOg was readily acted on by boiling 
water with formation of hydrogen chloride and a bromocamphanic 
acid identical with that obtained from the dibromoan hydride. It 
follows, therefore, that the bromo-acid in question is a ir-bromo-to- 
camphanic acid, and its constitution mny be expressed by one of the 
following stereoisomeric formulae. ^ 

Digitized by VjOOQIC 





III. v-Bromo-cw-i£^-camphaDic aoid. lY. ir<Bromo<^ra7W-i0-cainphanic aoid. 

Now it seems probable that the first of these two fornmlso repre* 
sents the configuration of the acid better than does the second, in spite 
of the fact that this view necessitates the assumption that intramole- 
cular change occurs in the formation of the acid from ntc-dibromo- 
camphoric anhydride ; when the properties of ordinary icMUtrnphanic 
acid are considered, it must be admitted that this compound is probably 
a eiff-lactonic acid, because the hydrozy-acid, of which it forms salts 
on treatment with alkalis, is net known in the free state, but immedi- 
ately passes into its lactone ; tiMiamphanic aqid, therefore, is doubtless 
stereochemically analogous to cia-Tr-camphanic acid, whereas the trans- 
iiMsamphanic acid, corresponding with the unstable ^raTM-ir^camphanic 
acid, has not yet been prepared. 

It is probable, then, that when lo-bromocamphorio anhydride is 
converted into to-camphanic acid, the carbozyl group changes its 
position stereochemically, just as it is known to do to a considerable 
extent in the reduction of t£^•bromooamphoric anhydride, a reaction 
which affords a mixture of ois- and trana-y or d- and <^iso-camphorio 
adds (Aschan, Ada Soe. Seimt.f&nn.^ 21, [v], p. 195). 

The conclusion thus arrived at, namely, that T-bromocamphanic 
aeid has the configuration represented by formula III, is in accord- 
ance with the behaviour of this substance, and also accounts for its 
conversion into hydroxy-ci«-«^-camphanic acid; when treated with 
potash, it may be supposed that it first gives a salt of to-hydroxy-ir- 
bromocamphoric acid, the tr-lactone ring undergoing hydrolysis, and 
that then potassium bromide is eliminated with formation of a different 
lactone ring, namely, that which is contained in ciairoamphanic 




T. vBromo-CM-te-KwnphaDic acid. YI. li^-Hydrozy-cM-r-camphanic acid. 

The last formula (YI) indicates the possible existence of a dUactane, 
which would be derived from hydroxy-cis-ir-camphanic acid by the 
eUmination of one molecule of water from the remaining -OH and 

Digitized by VjOOQIC 


-OOOH groups ; although the isolation of such a substance has not 
yet been accomplished, indications of its formation have been 
observed, and the neutral crystalline compound obtained by bromina- 
ting trontf-ir-camphaQic acid (Trans., 1896^ 60, 934), or the neutral 
oily product formed in the decomposition of iruMiibromocamphorio 
anhydride (see later), may possibly be a compound of this kind ; it is 
by no means improbable, however, that the existence of one lactone 
ring may, owing to stereochemical causes, hinder the formation of a 
second one, and thus render it difficult to obtain such a dilactone by 
ordinary methods. 

The farther oxidation of the compounds obtained from dibromo- 
camphoric anhydride has not yet been accomplished with satisfactory 
results in a single case; to give an instance of the difficulty of 
oxidising some of these products, it may be mentioned that a small 
quantity of the lactone of i&-hydroxycamphotricarboxylic acid may 
be heated at 100° with a mixture of concentrated nitric and hydro- 
chloric acids, or with a mixture of concentrated nitric, and sulphuric 
acids, and after several hours treatment, the lactone is deposited 
unchanged on keeping the solution at ordinary temperatures. 

Most of the compounds described in this paper are readily obtain- 
able in well-defined crystals, and the author has again to express his 
thanks to Mr. W. J. Pope for a number of interesting reports on the 
crystallographic characters of the various substances which have been 
submitted to him for examination. 

Part of the cost of this, and of all the other investigations on cam- 
phor derivatives, which have been published by the author, alone and 
in conjunction with Mr. W. J. Pope and Dr. A. Lapworth, has been 
defrayed by grants from the Koyal Society, for which the author desires 
to express his thanks to the €k)vemment Grant Oominittee. 


vuhDibramoeamphonc Anhydride^ CgHigBrg^^^^O. 

Dry 9r-bromocamphoric acid (Trans., 1896, 60, 924), which has been 
freed from a-nitro-air-dibromocamphor by washing with chloroform, 
is ground up with about one-tenth of its weight of amorphous phos- 
phorus, and gradually treated with bromine in a Wiirtz flask ; when 
the first^gorous action has subsided, the flask is heated on a water- 
bath and the addition of bromine continued very slowly, so that the 
quantity added in the course of about 3 hours is approximately twice 
the weight of the acid taken. During this operation, the -evolution 
of hydrogen bromide gradually slackens but without ceasing entirely, 
and bromine also escapes in small quanttti^^' ^ ^^^ ^^^ ^^^ ^^ ^^ 

ninitiypri H.i V llOOO I P 


normal reaction is difficult to recognise. The excess of bromine 
having been expelled, the product, which consists of a red, crystalline 
mass, saturated with a red oil, is well agitated with successive small 
quantities of cold water, and then with a little cold dilute alcohol ; 
iheee liquids remove most of the oily impurity, leaving a pale reddish 
or greenish crystalline product, the weight of which is rather greater 
than that of the original add, the average yield amounting to about 
90 per cent, of the theoretical. A small quantity of a crystalline bye- 
prodnct is obtained from the alcoholic washings, but the examination 
of this substance is not completed. 

The crude anhydride is very easily purified, without appreciable 
loss, by dissolving it in boiling chloroform, and precipitating the 
filtered solution with ether, repeating these operations if necessary ; 
a sample thus purified and dried over sulphuric acid was analysed. 

01779 gave 02307 CO, and 0-0643 Kfi. = 35-36; H = 401. 

01248 „ 0-1426 AgBr. Br » 47*8. 

O^oHi^rjOj requires = 35-29 ; H « 3-62 ; Br - 4706 per cent. 

Dibromocamphoric anhydride is readily soluble in boiling chloro- 
form, from which it crystallises in large, transparent plates (see 
below) melting at 209 — 210°*; it is also readily soluble in boiling 
eihylic acetate, acetic acid, and cold acetone, moderately in cold 
benxene, and sparingly in cold ether and alcohol It sublimes readily 
when heated in a test-tube, giving a solid sublimate of lustrous prisms. 
It seems not to be acted on by boiling quinoline, from which it separates 
again on cooling, but when heated with aniline it is vigorously 
attacked; it dissolves in warm, concentrated, sulphuric acid with 
evolution of hydrogen bromide, and on heating more strongly the 
solution darkens in colour considerably. 

Bailing water slowly converts the anhydride into bromocamphanic 
aeid and hydrogen bromide, and a boiling solution of half a molecular 
proportion of sodium carbonate in dilute alcohol brings about a 
OBiilar change ; boiling alcoholic or aqueous potash in eJccess causes 
the elimination of both the bromine atoms with formation of the 
potassinm salt of hydroxy-cwir-camphanic acid or dihydroxycamphoric 
add. Fusion with potash at a moderately low temperature also 
resnltB in the formation of a salt, from which u^hydroxy-cu^-cam- 
phanic acid is liberated on the addition of a mineral acid, but other 
producte appear to be formed in small quantities. When silver 
nitnte is added to a solution of the dibromo-anhydride in acetic acid, 
the separation of silver bromide soon commences, and after prolonged 
bmling io-hydroxy-«isirH»mphanic acid can be isolated from the isolu- 
ticm; the y-lactone of hydroxycamphotricarboxylic acid (Trans., 1896, 

* for coRoetunis to be applied to these qielting points, see Tians., 1697, 71, 968, 

Digitized by VjOOQIC 


60y 961) also seems to be produced under these conditions by the 
oxidation of the dihydroxycamphoric acid, which is probably formed 
as an intermediate product^ but the isolation of the lactone is not 
very easy. 

Dibromocamphoric anhydride is ladvorotatory. For the determina- 
tion of its specific rotation, a solution in chloroform of 1*158 grams 
was diluted to 25 cc, and examined at 14° in a 200 mm. tube ; the 
mean of 7 obseryations gave a= -2*9% from which [a]pa -31"2°.* 

The following is Mr. W. J. Pope's account of the crystals of the 
dibromo-anhydride, which were obtained by spontaneous evaporation 
of its solution in chloroform. 

'' ffti^-Dibromocamphoric anhydride crystallises in transparent^ rhom- 
boidal-shaped, orthorhombic plates (Fig. 1) possessing a calcite-like 
lustre. The dominant form is always the pinacoid a{100}, and the 
form ^{011} is usually the next largest ; these two forms give fairly 
good reflections. The prism |7{1 10} is, as a rule, much smaller than 
^{011}, and the pinacoid 5(010} is generally very small; the form 
r{101 } is always small but bright, and is frequently absent. 

Fio. 1. 

" Crystalline system. — Orthorhombic. 

a: 6 :c« 1*4844: 1:0-7083. 
« Forms observed.— a{ 100}, 6{010}, p{UO}, ^{011}, r{101}. 

'' The following measurements were obtained. 


Number of 






66'44'— 6617' 




67 81 — 68 12 

67 54 




88 80-- 84 24 

88 59 

88 58 

jrjr = 011:0lT 



109 1—109 40 

109 24 

109 28 


54 28 — 54 59 

54 41 

54 41 80" 



7014— 70 54 

70 87 




64 8—64 51 

64 27 

64 29 80 



50 47 — 51 86 

51 2 

51 1 

* The polfliimeter used in these determinations coidd only be read to 6', and 
the results may not be rery fu)cnr$t9i 

Digitized by VjOOQIC 

KiPFDra: DXRiYATnrn of camphoric acid, part in. 13S 

** The faces of the 3M>ne [001] are striated with lines parallel to the 
(; these lines are sufficiently well developed as frequently to 
disturb the measurements in that zone. The optic axial plane is 
0(001), and the azia-6 is a bisectrix of negative double refraction ; 
the optic axial angle is very large, but the h-hxis is probably the acute 
bisectrix* There is a fairly good cleavage on a{100}, and the cleavage 
faces are usually marked by the striations noted above ; it is note- 
worthy that the crystals are always tabular on a{100}. 

** After melting on a microscope slide under a cover slip in the usual 
way, the substance solidifies readily to a cubic modification ; when the 
plate cools to about 60°, this changes to a doubly refracting biaxial 
modification made up of large individual fragments which, as cooling 
eontinues, crack across perpendicularly to their long directions. The 
pieces are frequently perpendicular to an optic axis, and are marked 
by interlaced straight striations. The double refraction seems to be 
of negative sign, and the modification is in all probability with the 
crystals measured above." 

vuhDibromooamphorie Aeid^ CJB.i2'Br^{000B)^ 

It has been shown in previqus papers that when an anhydride 
cannot be converted into the corresponding acid by treatment with 
alkalis or with boiling water, owing to elimination of the elements of 
a halogen acid, as, for example, in the case of the anhydrides of 
«9-bromocamphoric acid (Trans., 1896, 69, 1), ir-bromooamphoric add 
(Trans., 1896, 69, 927), and ir-chlorocamphoric acid (Lapworth and 
Kipping, Trans., 1897> 71, 1), this conversion is easily accomplished 
by using concentrated nitric acid as the hydrolysing agent. This 
method can be applied with very satisfactory results for the prepara- 
tion of dilvomocamphoric acid from its anhydride. 

Dibromocamphoric anhydride dissolves readily in hot, concentrated 
nitric acid (sp. gr. 1*4), and if the solution be cooled after heating 
during a few minutes only, most of the anhydride is deposited 
uncfaaDged ; if, however, the nitric acid solution be heated on a water- 
bath in an evaporating dish, crystals of the dibromo-acid begin to 
separate after a short time, neither nitrous fumes nor bromine being 
evolved in any appreciable quantity. When most of the nitric acid 
has evaporated, a Uttle water is added to precipitate the rest of the 
dibromo-acid, and the colourless crystals are then separated- and dried 
in the air. This product is generally free from anhydride, but 
should the latter be present, it is easily removed by washing the 
erjstals with a little cold chloroform, in which the acid is insoluble, 
orneariy so. 

Digitized by VjOOQIC 


For analysiB, a sample was treated in this way, and then recrystal- 
Used from ether and dried at 100^ until constant in weight. 

01682 gave 02076 CO, and 0062i H^O. C = 3366 ; H = 412. 
CioHj^BrjjO^ requires C = 33-62 ; H - 391 per cent. 

mo-Dibromocamphorio acid crystallises from ether, in which it is 
very readily Boluble, in microscopic, fouiveided plates; it dissolves 
freely in cold acetone, ethylic acetate, and methylic alcohol, but is 
almost insoluble in benzene as well as in chloroform. It melts at 
about 210°, effervescing, owing to the escape of water vapour and 
hydrogen bromide, and becoming slightly brown, but like «o-bromo- 
camphoric acid and all the «^halogen derivatives of camphoric add, it 
is stable at 100°, and does not lose in weight even after having been 
heated during several hours. It is practically insoluble in boiling 
water, although just sufficiently soluble to impart to the water an 
acid reaction after heating for a few minutes ; it is probable, however, 
that the dibromo-acid does not dissolve unchanged, and that the 
acidity of the solution is due to hydrobromic and v-bromocamphanic 
acids, since these two compounds are rapidly produced when a solution 
of the dibromo-acid in dilute methylic alcohol is boiled. 

Hot concentrated nitric acid dissolves dibromocamphoric acid freely, 
and on cooling beautiful, transparent, flat, four-sided crystals axe 
deposited, but they are not large enough to be suitable for goniometric 
examination. The acid dissolves in dilute sodium carbonate, forming 
apparently the corresponding sodium salt, as on acidifying a freshly 
prepared solution with a mineral acid, dibromocamphoric add is repre- 
cipitated ; when, however, such a solution of the acid is boiled for a 
few minutes and then addified, crystals of ir-bromocamphanic add are 
depodted, hydrogen bromide having been eliminated. 

When dibromocamphoric acid is heated in a test-tube over a small 
flame, it first melts with effervescence and then distils, charring very 
slightly but evolving hydrogen bromide, and apparently traces of 
carbonic anhydride; the crystalline distillate consists of a mixture of 
vtr-dibromocamphoric anhydride and an add melting at 176°, whidi is 
doubtless ir-bromoeamphanic add. 

No attempts were made to prepare salts of the dibromo-add, on 
account of the readiness with which it is decomposed by water, alkali 
carbonates, and alkalis. It is interesting to note the relative stability 
of hot solutions of the acid in concentrated nitric acid and in water 
respectively ; whereas the former may be heated during several hours 
without suffering any appreciable change, dimination of hydrogen 
bromide takes place rapidly in the aqueous solutions ; this bd|i^yioi|r 
is similar to that of ir-bromocamphoric add, 

Digitized by VjOOQIC 


T-Bromo-to-Moroeamphorie Anhydride, OgHj^BrCl^^O^O, 

The object of preparing this compoand has already been stated ; the 
method was the following. Dry powdered ir-bromocamphoric acid was 
treated with a slight excess of the theoretical quantity of phosphorus 
pentachloridei the colourless liquid product heated on a water-bath, 
and dry chlorine slowly bubbled into it for about 4 hours, or until the 
evolution of hydrogen chloride almost ceased. The pale yellow oil 
thus obtained was gradually treated with ice cold water, whereon it 
quickly solidified, and was then purified by washing with water and 
dilute alcohol successively ; for analysis, a sample was recrystallised 
twice from chloroform and then dried over sulphuric acid. 

01576 gave 0-2349 CO, and 00602 HjO. C = 40-66 ; H = 4-24. 
CjoHjs^lBrO^ requires C » 40*62 ; H « 406 per cent. 

s^Bromo-i9-ehlorocamphoric anhydride, like the corresponding di- 
bromo-compound, crystaJlises best from chloroform, from which it is 
deposited in lustrous prisms (see later) melting at 214 — 216^, or 6^ 
higher than its analogue ; a mixture of the dibromo- and chlorobromo- 
anhydrides shows no sign of melting until the temperature rises 
to 209 — 210° — the melting point of the former — an indication of 
the isomorphism of the two compounds which was confirmed by 
the crystallographic examination made by Mr. W. J. Pope (see 
p. 136). 

In most respects, the properties of s-bromo-u^-chlorocamphoric anhy- 
dride are so similar to those of the corresponding dibromo-compound 
that further description is unnecessary, but one rather interesting 
difference in behaviour may perhaps be noted. When the chlorobromo- 
eompound is crystallised from a mixture of chloroform and ether, it is 
sometimes deposited in long, slender prisms, or needles, as well as in 
the compact prisms already referred to ; these two kinds of crystals 
differ, not only in appearance, but also in behaviour, as the former 
become opaque when kept over sulphuric acid or heated on a water- 
bath, whereas the latter remain transparent ; both forms, however, 
melt at 214— 215^ 

Experiments showed that the needles are formed when the crystal- 
lisation of the ethereal chloroform solution takes place at low 
temperatures, as, for example, at - 6°, whereas, when the solution 
is kept at about the ordinary atmospheric temperature, one or other, 
or both forms may be deposited ; ethereal chloroform solutions of the 
dibromo-anhydride, crystallised at temperatures just below 0°, did not 
deposit anything but the rhombic prisms described above. 

Digitized by VjOOQIC 


This anhydride, like the corresponding dibromo-compound, is l»vo- 
rotatory ; a solution of 1*1 5«3 grams in chloroform dilated to 25 cc, 
and examined at 14^ in a 200 mm. tube, gave a » - 2*4° as the average 
of seven observations : hence [a]D== - 26*1°.* 

" irt/7-Chlorobromocamphoric anhydride crystallises in large, trans- 
parent, orthorhombic prisms (Fig. 2), which appear rather more lostrotts 
than the crystals of the corresponding dibromo-compound. The 
dominant form, as in the latter case, is usually the pinaooid a{100}, 
but the form j^jlOO} is sometimes the most developed ; the pinacoid 
5{010} is also usually well developed, and ^{011} is generally broader 
than in dibromocamphoric anhydride. The crystals of the latter 
give much better results on measurement than do those of the com- 
pound now described. 

Fig. 2. 





yT-- ! 









•* Crystalline system. — Orthorhombic. 

a :6:c = 1-4789 :1:0-7107. 
"Forms observed : a{100}, 6{010}, je){110}, ^{011}, r{101}. 

'' The following measurements were obtained. 


Number of 




W= Oil: Oil 


65'' 1'— 66'64' 
07 49— 69 2 
33 20 — 34 48 
108 42 -109 67 
63 89— 66 10 
69 67 — 71 28 
63 19—66 6 

68 31 
34 7 
109 16 
64 38 
70 48 
64 24 

68** 8' 

34 4 

109 12 

.64 S« 

64 20 

'< The forms in the zone [001], and more especially the pinacoid 
a{100}, ave^'^arked with striations parallel to the c-axis. The optic 
axial plane is c{001}, and the axis h is the acute bisectrix ; the double 

* Compare footnote, p, 182. 

Digitized by VjOOQIC 


refraction is negative in sign, and the optic axial angle is large. 
There is a very poor conchoidal cleavage on a{100}. 

''After melting on a microscope slide under a cover slip, the com- 
pound solidifies readily, yielding a singly-refracting cubic modification ; 
this, before it cools to the atmospheric temperature, changes to a 
biaxial doubly-refracting modification, which is formed in large plates 
very similar to those constituting the film of the corresponding 
dibromo-derivative. These plates are marked with interlaced, straight 
striations, and are frequently perpendicular to an acute bisectrix of 
large axial angle and of negative double refraction. 

'* As would be expected from the chemical relationship between the 
two compounds, uto-dibromo- and iruHshlorobromo-camphoric anhydride 
are isomorphous; the axial ratios are of the same order and cor- 
responding angles, as shown in the following table, do not differ 
greatly in the two cases. 







+ 0-0055 








+ 0-0149 




+ 0*6' 


33 58 

84 4 



35 18 30" 

35 24 



54 41 80 

64 36 

+ 5 80 


64 20 80 

64 20 

+ 9 30 


25 30 30 

25 40 


" It will be seen that the dimensions measured in the zone [001] 
differ least in the two cases, whilst those in the parametral zone [010] 
exhibit the greatest differences. The same forms are present on 
crystals of both anhydrides and striations parallel to the zone axis 
observed in the zone [001] are also observed in both cases. 

*' Amongst the differences may be noted that the crystals of the 
dibromo-compound show a much better cleavage parallel to {100} than 
do those of the chlorobromonlerivative ; again, the crystals of the 
former are really much better developed and give more trustworthy 
measurements than those of the latter substance. Further, there is a 
well-marked difference in habit. Crystals of the dibromo-anhydride 
are always tabular, whilst those of the chlorobromo-derivative are 
much more prismatic in habit ; this is mainly due to a large develop- 
ment of the forms p{110} and 5(010} on the latter crystals. 

^The behaviour of the two substances after melting and solidification 
is, in accordance with the isomorphism, extremely similar." 

Digitized by VjOOQIC 


v-BronuHv-Maroeampharic Acid, C^lB.i2^rCl{(X)OR)^ 

This acid can be prepared from its anhydride by the same method 
as that employed in the case of the dibromo-anhydride, namely, by 
dissolving the anhydride in concentrated nitric acid (sp. gr. 1*4), and 
then evaporating the solution on the water-bath. The colourless 
crystals of the acid are dried on porous earthenware, washed with a 
little chloroform to remove traces of unchanged anhydride which may 
be present, and then dried at 100^; for the analysis, a sample was re- 
crystallised from ether, as the crystals retain small quantities of 
occluded nitric acid, which is not easily expelled even at 100^. 

01689 gave 0*2395 CO, and 0*0693 H^O. » 38*67 ; H - 4 56. 
CioH^^BrOlO^ requires 0* 3829 ; H = 4*46 per cent. 

ir-BromoH(^-chlorocamphoric acid is a microcrystalline, colourless 
powder, as prepared in the above manner ; it melts at about 197^ when 
heated fairly slowly from about 80^, effervescing vigorously, but as 
this temperature is a decomposition rather than a melting point, it 
varies a little with the rapidity of heating. There is, however, a much 
larger difference between the melting points of the bromochloro- and 
dibromo-acids, than between those of the corresponding anhydrides. 
In all ordinary properties, the bromochloro-acid resembles the dibromo- 
compound ; it is, however, rather more soluble in boiling water, from 
which it separates in long, transparent prisms, which seem to contain 
water of crystallisation, as they become opaque when heated and melt 
at about 197°. It separates from cold concentrated nitric acid in 
nodular, opaque masses, and from ethereal chloroform in colourless, 
transparent, well-defined microscopic prisms ; its behaviour towards 
boiling water is referred to later. 

v-Bromoeamphanic Add, CXJOH-CgHi^Br^^Y . 

Various method^ can be adopted for preparing this compound from 
dibromocaitphoric anhydride, or from the corresponding acid, some 
of which l^ve already been mentioned. It may be obtained in large 
quantities by boiling the finely divided anhydride with a large 
volume of water, but as this process occupies a long time, owing to 
the slight solubility of the anhydride, the following method is more 

The anhydride is dissolved in a small quantity of boiling acetic 
acid, and after adding a little water until a turbidity is just produced 
in the hot solution, the latter is heated on a sand-bath, more water 

Digitized by VjOOQIC 


being added from time to time as the anhydride decomposes ; daring 
this prooess, the solution almost invariably acquires a distinct bright 
pink colour^ which slowly changes to brown, the liquid then depositing 
traoee of some tarry matter, doubtless due to imparity in the (crude) 
anhydride used. The hot solution is finally filtered, and on cooling 
the T-bromocamphanic add is deposited in almost colourless needles ; 
further quantities are obtained on evaporating the mother liquors, the 
yield being practically theoreticaL 

For analysis, a sample was purified by recrystallisation from a 
mixture of ethylic acetate and chloroform and dried at 100^. 

01676 gave 0-2518002 and 00697 H,0. C -43-57; H = 4-91. 
OioHijBrO^ requires C = 43-32 ; H = 4-69 per cent. 

v-Bromocamphanic acid usually separates from hot water, in which 
it is only moderately easily soluble, in small, well-defined, lustrous 
prismSy which do not lose in weight at 100^. It is comparatively 
sparingly soluble in boiling benzene, and only moderately so in boiling 
chloroform, but it dissolves readily in acetic acid, alcohol, and ethylic 
acetate ; it crystallises best from cold, dilute acetic acid, from which 
it iA deposited in transparent prisms described later. It melts at 
176 ^tI27^ without decomposing, but when heated at its boiling point it 
darkens considerably and seems to give off a little hydrogen bromide ; 
the distillate soon crystallises on cooling, and consists of slightly im- 
pure bromocamphanic acid. 

v^Bromocamphanic acid is a very stable substance in many respects, 
and boiling nitric acid, dilute or concentrated, does not oxidise any 
appreciable quantity of it in the course of a few hours. Prolonged 
b(^ing with a solution of chromic anhydride and dilute solphuric 
acid seems to result simply in the elimination of hydrogen bromide, 
the bromocamphanio acid being converted into hydroxy-ci^-Tr-camphanic 
acid ; the last-named compound is also formed when the bromo-acid is 
boiled with an aqueous solution of silver nitrate, but other substances 
are also produced. When the acid is treated with ammonia under 
suitable conditions, a product is obtained which is almost insoluble in 
water, but which has not yet been analysed or examined. 

Mr. Pope's description of this substance is as follows. " The crystals 
of bromocamphanic acid are very beautiful, transparent, orthorhombic 
prisms (Fig. 3), which have a very brilliant, glassy lustre. The prism 
|>{110} is always dominant, but gives poor results on measurement 
owing to the presence of vicinal faces and of striations parallel to the 
c-axis; the forms ^{011} and r{101} are next in size, r{ 101} being 
usually rather the larger, and give very brilliant reflections, so that 
the mMHrarements obtained from them may be relied upon. The 
piBaooid c{001} is very «mall and frequently absent. 

Digitized by VjOOQIC 


Fio. 8. 

" Crjstallme System. — Orthorhombic. 

a: 5 :c»l-4654: 1:0-9501. 
"Forms observed.— c{001}, />{110}, ^{011}, r{lOl}. 
** The following angular measurements were obtained. 

Number of 








86n7'— 86'69' 



rr* 101:101 


71 12— 71 26 



rr= 101:101 


108 84—108 48 

108 40 

108 41 8 



70 41— 70 56 

70 49 




55 49— 56 16 

56 59 

56 57 



58 5— 58 24 

58 15 

53 14 



86 52— 87 20 

87 4 

87 8 



92 87— 98 11 

92 65 

92 52 



46 1— 46 54 

46 21 

46 26 



110 47—11158 





68 20— 69 18 

68 47 

68 87 

" The plane c(OOl) is the optic axial plane, and an optic axis emerges 
obliquely through each face of the form /?{ 110} ; the double refraction 
is strong and the optic axial dispersion slight. After melting, the 
substance solidi^es readily, and is thus obtained in large individual 
flakes, which, are often marked by strin crossing each other at about 
60^, and cracked across their loiTgest dimension on cooling ; the flakes 
are frequently perpendicular to an optically negative bisectrix of a 
large axial angle." 

Methylic fr-Bromoeamphanate, COOMe*OgHijBr^_;[Y. 

This ethereal salt is conveniently prepared by passing hydrogen 
chloride into a solution of the acid in methylic alcohol and then 

Digitized by VjOOQ IC 


evaporating at the ordinary temperatare ; the product, which is slowly 
deposited in felted needles, can be purified by recrystallisation.from a 
mixture of ether and chloroform. 

01662 gave 02777 COj and 00832 HaO. C=-45-67 ; H-6-56. 
CjiHisBrO^ requires C =» 45*36 ; H = 616 per cent. 

Methylic ir-bromooamphanate crystallises from ether and from most 
other solvents in long, transparent needles melting at 87 — 88° ; it is 
very readily soluble in cold chloroform, methylic alcohol, and boiling 
ether, but comparatively sparingly soluble in light petroleum. Massive 
crystals are easily obtained from a solution in a mixture of ether and 
ohlorolbrm, and a specimen of the compound thus obtained was 
examined by Mr. W. J. Pope, whose report is now given. 

*'The crystals of methylic bromooamphanate are transparent ortho- 
rhombic plates or prisms, possessing a high lustre (Fig. 4) ; the habit 
of the crystals varies considerably, those of prismatic habit are 
lengthened in the direction of the axis-a, and somewhat flattened in 
that of the 5-axis, the dominant forms being a{100} and r{101} ; the 
tabular crystals are flattened on two parallel faces of the form r{101}. 
Yery good results are obtained on measuring the crystals. 

Fio. 4. 

"CJ^stalline System. — Orthorhombic. 

« Forms observed,— a{ 100}, ft{010}, ^{110}, r{lbl}. 
''The following angular measurements were obtained. 

Kamber of 
1 nients. 



42 18 
59 8 
69 45 


rri= 101:101 


47*84'— 47'47' 
42 12—42 27 
59 0—59 28 
61 81 —61 60 
69 89 —69 52 


61 44 
69 47 


Digitized by VjOOQIC 


'^ There is a poor oonchoidal cleavage parallel to 5{010}, and the axi8-5 
is the acute bisectrix ; the optic axial plane is c(001)y and the double 
refraction is positive in sign, and strong. The optic axial angle is 
fairly large and the optic axial dispersion is slight. 

*' The substance solidifies readily after melting on a microscope slide 
under a cover slip, giving large, individual flakes ; most of these are 
nearly perpendicular to an optically negative bisectrix of a very large 
optic axial angle, and these are full of symmetrically arranged egg- 
shaped bubbles and of intersecting striations. A few fragments are. 
usually to be observed which show no bubbles or strie ; these are 
nearly perpendicular to the optically positive bisectrix of a fairly large 
optic axial angle. This modification is doubtless identical with the 
crystals measured above." 

Methylic ir-bromocamphanate is slowly attacked by concentrated 
aqueous ammonia at ordinary temperatures, and is thereby converted 
into a crystalline substance, which, judging by its melting point and 
other ordinary properties, is identical. with the «--bromocamphanamide 
produced by the action of aqueous ammonia on trtr-dibromocamphoric 
anhydride. This compound may now be described. 

vBramotO'camphomamidef NHg'CO'CgHijBr^Y. 

When finely divided iruvdibromocamphoric anhydride is left in 
contact with concentrated aqueous ammonia at ordinary temperatures, 
it slowly changes into a rather more bulky mass of small prisms, but 
without passing into solution to any noticeable extent ; after keeping 
the mixture for about two days, the product is separated by filtration 
from the ammoniacal solution, which contains a small quantity of a 
readily soluble, crystalline, ammonium salt, washed with cold water, 
and recrystallised from dilute methylic alcohol. The substance 
obtained in this way is the amide of 1l^bromorU^'Camphanic acid, and 
an analysis of it gave the following result. 

0-1678 gave 0*2671 COj, and 0-0796 HgO. = 43-41 ; H = 5-27. 
CjoHi^OgBrN requires C«43-47; H = 507 per cent. 

The nature of this compound is further established by the fact 
already mentioned, namely, that it is formed on treating methylic 
ir-bromocamphanate with aqueous ammonia, and also by its behaviour 
on hydrolysis ; when boiled for a short time with concentrated hydro- 
chloric acid, it is converted into a crystalline acid, which melts at 
176 — 177^, and has all the properties of ir-bromocamphanic acid. 

ir-Bromo-u^camphanamide crystallises from most solvents in lustrous, 
transparent needles or prisms melting at 161 — 162^ ; it is very readily 

Digitized by VjOOQIC 


soluble in cold chloroform, acetone, ethylic acetate, acetic acid, and 
most other solvents, and it also dissolves freely in boiling water, bat 
it is insoluble, or nearly so, in a cold dilute solution of sodium car- 
bonate. It seems to be dimorphous, as the transparent crystals 
deposited from dilute methylic alcohol become opaque at about 140^ 
when slowly heated, and do so at even lower temperatures when the 
tube containing them is rubbed gently. 

FamuUian qfTr-Bromocamphanic Acid from v-BronKhto-chlorocamphorie 
Anhf^dride. — It was stated in the introduction that ir-bromo-fo-chloro- 
camphoric anhydride is decomposed by boiling water, giving hydrogen 
chloride, and an acid identical with the ir-bromo-tiM^amphanic acid, 
prepared in a similar manner from 1rt(^^ibromocamphoric anhydride ; 
this statement rests on the following experimental evidence. 

When bromochlorocamphoric anhydride is boiled for some hours 
with diluted acetic acid, and the filtered solution then allowed to cool, 
a substance crystallising in colourless prisms is deposited ; this com- 
pound, after having been purified, melted at 176 — 177^, and in appear- 
ance and other properties seemed to be identical with ir-bromo- 
camphanic acid. The great similarity between dibromo- and bromo- 
chloro-camphoric anhydrides, however, if repeated in the case of the 
bromo- and chloro-camphanic acids, might render the distinction 
between the two latter a matter of some difficulty ; for this reason, it 
was necessary to make the following analysis, the results of which 
show that the decomposition product of the bromochloro-anhydride is 
really a v-bromocamphanic acid. 

0-1672 gave 0-2621 00, and 00679 H^O. = 43-74 ; H = 4*80. 
OioHjjBrO^ requires = 43-32 j H = 4-69 per cent. 

ffydroocf/^is-W'Camphamc Acid, 

It has been shown in earlier papers that ets-ir-camphanic acid is 
slowly oxidised by potassium permanganate in alkaline solution, being 
converted into a hydroxy-c{»-7r-camphanic acid, which is the lactone of 
a dihydroxycamphoric acid; this hydroxy-acid can be obtained in 
various ways from dibromocamphoric anhydride and its derivatives. 

When, for example, dibromocamphoric anhydride is boiled^or some 
time with alcoholic potash, or fused with potash at a moderate 
temperature, both the bromine atoms are removed, and hydroxy-oM- 
ir-camphanic acid can be isolated from the product by methods which 
it is unnecessary to describe. Again, when ir-bromocamphanic acid is 
heated with silver nitrate in aqueous solution, silver bromide is 
rapidly deposited, and an acid having all the properties of hydroxy- 
ei9-s^<:amphanic acid can be obtained from the solution. 

A number of experiments were made in the hope of isolating a 

Digitized byVjOOQlC 


dihydroxjcamphoric acid, or the corresponding dilactone; for this 
purpose, the dibromo-anhydride, dibromo-acid, and ir-bromocamphanic 
acid, were separately treated with aqueous silver nitrate under various 
conditions ; in nearly every case the product seemed to be a mixture 
of two or three organic compounds, and it was not easily separable 
into its components by fractional crystallisation or by other methods ; 
hydroxy-cM-^r-camphanic acid was isolated in almost every instance, 
and also a crystalline substance, which from its melting point and 
other properties was found to be identical with the y-lactone of 
hydroxycamphotricarboxylic acid. Further, in several experiments, 
small quantities of an oily product were obtained ; this substance, 
when purified, was only sparingly soluble in boiling water (the other 
two products are readily soluble), and apparently insoluble in cold 
dilute sodium carbonate ; it was vigorously attacked by hot concen- 
trated nitric acid, being converted into the y-lactoneof camphotricarb- 
oxylio acid. These observations seem to show the existence of a 
dilactone of dihydroxycamphoric acid amongst the products of the 
action of silver nitrate on the bromo- and dibromo-compounds in 

Hydroxy-ots-ir-camphanio acid is prepared far more easily by one of 
these methods than by the oxidation of ct«-ir-camphanio acid, and its 
investigation, as well as that of some of the other compounds related 
to it, is being continued ; the crystals of this acid, obtained from an 
acetone solution, are of large size and suitable for goniometrio 

Univebsity Gollbob, 

XVII. — a-Ketotetrahydronaphthalene. 

By Fbedkbio Stanley Eifpino, Ph.D., D.Sc., F.R.S., and Alfbxd 


The conversion of phenylpropionic chloride into a-hydrindone by the 
action of aluminium chloride under the conditions described by one of 
us in a previous communication (Kipping, Trans., 1894, 65, 680) is 
an example of intramolecular condensation by which an open- is- 
oonverted into a closed-carbon chain ; the reaction also provides by 
far the most convenient method yet known for the preparation of 
a-hydrindone, as, with due care, the yield is invariably very good, and 
phenylpropionic chloride itself can be obtained without any difficulty 
in almost unlimited quantity ; this method of preparation was, there- 
fore, made use of in later investigations of the derivatives of a-hydrin- 
done (Kipping and Revis, Trans., 1897, 71, 238). 

Digitized by VjOOQIC 


That such a reaction would be capable of more general application 
wa8| of course, extremely probable, bat until quite recently no oppor- 
tunity offered itself of putting this view to the test of experiment ; 
now, however, the reaction is being tried with various substances, 
and in the present paper some of the first results of this work are 

Taking the well-known synthesis of a-naphthol from phenyl-^y- 
crotonic acid by Fittig and Erdmann as an indication of the ease with 
which intramolecular condensation may occur in the case of a benzene 
derivative containing a suitable unsaturated side chain, it seemed very 
probable that a similar change might be brought about in the case of 
{^enylbutyric acid, if the chloride of this acid were treated with 
alaminium chloride in a suitable manner ; if so, the analogy between 
phenylpropionic and phenylbutyric acid would be complete, the one 
giving a-hydrindone, the other a-ketotetrahydronaphthalene. 

Phenylbatyrio chloride. a-Eetotetrahydronaphthalene. 

Experiments showed that a-ketotetrahydronaphthalene is, in fact, 
formed bj the action of alaminium chloride on phenylbutyric chloride 
under the oonditions described below, but, so far, satisfactory yields 
have not been obtained, owing probably to the readiness with which 
the ketone, when once formed, undergoes further condensation of the 
ordinary type ; a sufficient quantity of this substance, however, has 
been obtained to enable us to study its properties in some measure, 
and to prepare from it various crystalline derivatives. 

orKetotetrahydronaphthalene is isomeric with, and closely related 
to, the /3-keto-compound, which was first prepared by Bamberger and 
Lodter {Ber.y 1893, 26, 1833) by the action of alkalis on '< tetrahjdro- 
naphthylene chlorhydrin" (2 :^ 3-chlorhydroxytetrahydronaphthalene), 
and afterwards examined by Bamberger and Yoss {Ber.f 1894, 27, 
1547). As^ far as can be judged, the two compounds resemble one 
another very closely in properties, and both show the ordinary general 
reactions of ketones ; the semicarbazone, oxime, hydrazone, and para- 
bromhydraxone of the a-keto-compound are described in this paper, 
the oxime and hydrazone of the ^-ketone having been previously 
prepared by Bamberger and Voss {he. cit,). 

That the substance we describe as a-ketotetrahydronaphthalene really 
has this constitution is shown, not only by its method of formation 
and ketonic properties, but also by the fact that it can be indirectly 
converted into a base which seems to be identical with oc-tetrahydro- 

Digitized by VjOOQIC 


a-naphthylamine ; this base was first obtained by Bamberger and 
Bammann {Ber., 1889, 22, 951) by reducing tetrahydro-1 : 6-naphthyl- 
enediamine with sodiom and amylic alcohol, and then eliminating tiie 
amido-group combined with the unreduced aromatic nucleus. 

^2 ^ /\^0H, 

JCH, \/\/^^ 

1 : 5-Tetrahy(lrooaphtbyl- Tetraliydro-a naphthyl- a-Ketotetrahydro- 

enediamine. amine. naphthalene. 

The conversion of a-ketotetrahydronaphthalene into this base was 
accomplished by first preparing the oxime, and then reducing the 
latter with sodium amalgam in acetic acid solution. 

The study of this interesting ketone is being continued. 

A portion of the expense incurred in this and in two earlier inves- 
tigations of a similar kind (Kipping, Trans., 1894, 65, 680; Kipping 
and Eevis, Trans., 1897, 71, 238) has been met by a grant .kindly 
awarded by the Oovernment Grant Oommittee of the Royal Society. 


Prepa/ration qf PhenyUmtyrie Add, — ^The preparation of a large 
quantity of phenylbutyric acid by the method which we employed is 
not a very easy task, but we were unable to find any process which 
appeared to be more suitable, recorded in the literature. 

Starting with benzaldehyde, anhydrous sodium succinate, and acetic 
anhydride, we first prepared the mixture of phenyl-)9y-crotonic and 
phenylparaconio acids under the conditions laid down by Jayne 
{Anruden, 1883, 216, 97), but instead of then separating these 
two compounds by dissolving out the phenyl-)3y-crotonic acid with 
carbon bisulphide, we submitted the dried mixture directly to dry 
distillation under reduced pressure (about 30 mm.), whereby, as 
Jayne has already shown {loo. eU.), the phenylparaconic acid is 
decomposed into phenyl-)3y-crotonic acid • and carbonic anhydride, a 
small quantity of phenylbutyrolactone being also formed from the 
phenyl-^y-crotonic acid by an intramolecular change. The distillate, 
which solidifies on cooling to a brown, pasty, crystalline mass, was 
spread on porous earthenware to separate the phenylbutyrolactone 
and other oily impurities, and afterwards recrystallised from hot 
carbon bisulphide, from which the phenyl-jSy^rotonic acid was deposited 
in almost colourless needles ; the last portions of this acid generally 
contain a little benzoic acid, this compound being always present in 
the original condensation product in large quantities ; it is best 

Digitized by VjOOQIC 


removed by heating the mixture on a water-bath, the residue being 
again crystallised from carbon bisulphide. 

The phenyl-^y-crotonic acid was thus obtained in colourless needles 
melting sharply at 86^, but the yield was always very poor, owing to 
the formation of tarry matter in the original condensation process. 

For the conversion of this unsaturated compound into phenylbutyrio 
acid, the pure acid was reduced with sodium amalgam in alkaline 
solution, and after precipitating with hydrochloric acid and filtering, 
the rest of the product was extracted with ether, and the whole 
purified by recrystallisation from this solvent ; the acid then melted 
at 47 — 48° as stated by Jayne {toe. cit.). We did not experience any 
difficulty in reducing the phenyl-)3y-crotQnic acid at ordinary tempera- 
tures, although, according to Jayne, prolonged treatment and warming 
are necessary. 

Phenylbutyric chloride, OgHg-CHj-OHj-OHj-COCl, is easily prepared 
by treating the dry, powdered acid with a slight excess of the theoreti- 
cal quantity of phosphorus pentachloride at ordinary temperatures, a 
vigorous action immediately setting in ; instead of separating the 
product from the phosphorus oxychloride by fractional distillation 
under reduced pressure, as in the case of phenylpropionic chloride 
(Trans., 1894, QQ, 484), we merely heated the oil on a water-bath under 
diminished pressure until the oxychloride had volatilised, using the 
residual phenylbutyric chloride, which is sufficiently free from impurity, 
for the subsequent experiments. 

Phenylbutyric chloride, as thus obtained, is a yellowish, mobile 
liquid, having a slight, but not very unpleasant, odour ; when poured 
into water, it rapidly solidifies to crystals, which are only very slowly 
decomposed by cold water, and which do not immediately dissolve in 
a cold dilute solution of sodium carbonate. 

Action of Aluminium Chloride on Fhenyllmtyric Chloride. 

Experiments were made in order to ascertain the conditions under 
which the conversion of phenylbutyric chloride into a-ketotetrahydro- 
naphthalene could be accomplished, and we commenced by employing 
very much the same method as in the case of a-hydrindone, the acid 
chloride, mixed with about four times its weight of light petroleum 
(b. p. 40 — 60°), being treated with its own weight of anhydrous alu- 
minium chloride. Under these conditions, however, only a very 
slight reaction, if any, occurred, and even after heating on the water- 
bath during 30 — 40 minutes, the phenylbutyric chloride was found to 
be unchanged ; on employing as diluent a sample of light petroleum 
boiling at 60 — 80°, the reaction was almost as sluggish as before, but 
with petroleum boiling at 100 — 110° a vigorous evolution of hydrogen 

Digitized by 


14A KIPPINQ ANt> diLt : a-&EtOT£T^MYl>R0KAl»HTHAtEN& 

chloride set in on heating, and in a short time the reaction abruptly 
ceased. An examination of the product from this last experiment 
showed that the yield of ketone was extremely small, most of the add 
chloride having been converted into a pale brown resin. For this 
reason, a smaller proportion of aluminium chloride was employed in 
the later experiments, and the results appeared to be distinctly better. 

The most satisfactory yield, so far, has been obtained by working 
in the following manner. Fhenylbutyric chloride (5 parts) is dissolved 
in light petroleum (26 parts) boiling at 100 — 110°, aluminium chloride 
(4 parts) is added, and the mixture is rapidly heated on a boiling water 
bath in a flask provided with reflux condenser ; after a very short 
time, hydrogen chloride is very rapidly evolved, and during this 
reaction the flask is repeatedly and vigorously shaken. At the end of 
about 10 minutes, or when the evolution of gas suddenly ceases, the 
contents of the flask are cooled and water carefully added; finally, the 
mixture is submitted to distillation in steam. The petroleum which 
passes over first contains only a small quantity of the ketone, the rest 
distilling over slowly with the steam and leaving in the flask a brown, 
oily, non-volatile substance, which, on cooUng, solidifies to a brittle 
resin; the ketone is isolated by evaporating the petroleum and by 
extracting the aqueous portion of the distillate with ether. 

The a-ketotetrahydronaphthalene obtained in this way as a pale 
yellow oil is probably slightly impure, and is further treated in the 
manner described below ; the yield is by no means satisfactory, being 
at the most only about 10 per cent, of the theoretical, but it is 
probable that, as in the case of a-hydrindone, further experiments will 
lead to the discovery of some slight modification of the process which 
will give much better results. 

The principal product of the reaction is the brown, brittle resin, 
referred to above ; this substance is probably formed from a-ketotetra- 
hydronaphthalene by a condensation process similar to that which 
occurs in the case of '< anhydrobishydrindone," or by the further 
action of phenylbutyric chloride on the ketone in presence of 
aluminium chloride. To obtain some information on this point, some 
of the resinous substance was oxidised with nitric add, and it was 
found to be ultimately converted into phthalic acid; this fact seems 
to justify the above assumptions, and to indicate that if this secondary 
reaction could be prevented the yield of the ketone would be as good 
as in the case of a-hydrindone. 

a-EeioMrahydr(maplUhalene, OeH4<\^^i^ 

'When working with small quantities of a-ketotetrahydronaphthalene, 
the best method of purification is doubtless the following. The crude 

Digitized by VjOOQIC 


oil ia oonyerted into its sparingly soluble semicarbazone in the manner 
described later, and after recrjstallisation it is heated in a small 
WfirtE flask with rather more moderately-concentrated hydrochloric 
acid than is required to combine with the semicarbazide ; the semi- 
carbazone is thus decomposed with regeneration of the ketone, which 
can be distilled off in a current of steam and extracted with ether. 
The addition of a little acetic acid, in which the semicarbazone is 
readily soluble, hastens the reaction. In this operation, there is no 
appreciable charring if the pure semicarbazone be employed, and the 
deoompodtion appears to be normal, as expressed by the following 

a-Ketotetrahydronaphthalene is a colourless, mobile, highly refrac- 
tive liquid. It does not crystallise when kept for some days at 
ordinary temperatures, or when cooled to 0°, whereas the correspond- 
ing /3-ketone solidifies when cooled, and melts again at 18^ (Bamberger 
and Yoss, loe. eU.) ; it does not seem probable that this difference in 
behaviour is due to the presence of impurity in the a-ketone, as the 
prooees just given appears to be a very satisfactory method of purifi- 

orKetotetrahydronaphthaleoe is specifically heavier than water at 15^, 
and ifl moderately easily volatile in steam. It has only a faint odour, 
recalling that of camphor, but when warmed it has a distinct odour 
of peppermint. It shows many of the ordinary reactions of a ketone, 
and yields crystalline products with hydroxylamine, phenylhydrazine, 
^ ; it does not appear to dissolve in, or to combine with, sodium 
hydrogen sulphite in aqueous solution, although the )9-ketone forms a 
crystalline additive product with this reagent. 

The ketone itself was not analysed, as its composition is established 
by its method of formation and properties, and by the analysis of the 

a-EetaMrahpdronaphthaJene Semtcarbcuane, CiqHjqIN*NH*C0*KH2. 

When the crude ketone, obtained by the method described above, 
is heated with semicarbazide hydrochloride and sodium aceta^ in 
aqueous alcoholic solution, the separation of an almost colourless, 
crystalline compound soon commences, and the reaction is completed 
by warming on the water-bath during about 2 hours; the hot 
solution is then diluted with water, allowed to cool, and the product 
separated by filtration, and washed well with cold water. 

When purified by recrystallisation from hot alcohol and dried over 
sulphuric acid, it gave the following result. 

Digitized by VjOOQIC 


0-1640 gave 03885 00^ and 00968 HjO. C = 646 ; H = 66. 
OjiHigNgO requires « 65'0 ; H = 6*4 per cent. 

a-Ketotetrahydronaphthalene semicarbazone crystallises from alcohol 
in long, transparent needles or prisms, usually forming aggregates of a 
rosette-like form ; these crystals are distinctly yellow, their colour being 
almost as intense as that of quinone, but when in a fine state of division 
the substance appears almost colourless. The melting point is not very 
definite, for when heated moderately quickly the finely-divided sub- 
stance melts at about 217?, but larger crystals only sinter at this 
temperature, and do not liquefy completely until about 220% when 
effervescence sets in and the substance darkens slightly ; on heating 
more strongly, a large quantity of gas is disengaged and a yellow liquid 

The semicarbazone is comparatively sparingly soluble in boiling 
chloroform and ethyUc acetate, and apparently insoluble in water ; it 
dissolves fairly easily in boiling alcohol and boiling acetone, and also 
in warm acetic acid, but it seems to be decomposed on boiling its 
acetic acid solution. A.s stated above, this compound may be con- 
veniently employed in the purification of the ketone, as the latter ia 
immediately regenerated on warming the semicarbazone with hydro- 
chloric acid ; it also affords the best means of detecting and identifying 
the ketone, the phenylhydrazone being far less suitable for such 

a-Ketotetrahydranaphthalene Phenylhydrazone, CiQHjoIN'NHPh. 

This compound is easily obtained by treating the purified ketone 
with phenylhydrazine acetate in dilute acetic acid solution in the 
usual manner, combination taking place spontaneously ; after warm- 
ing gently, water is added, and the product, which is precipitated as 
a thick, yellow oil, is washed well with cold water, and then dissolved 
in methylic alcohol. From this solution, the hydrazone separates, on 
spontaneous evaporation, in a crystalline condition, but if the warm 
solution be rapidly cooled, the compound is generally deposited as an 
oiL The crystals obtained from methylic alcohol and other solventa 
are massive, transparent, almost colourless six-sided and rhomboidal 
plates melting at 84 — 85^, at the same time effervescing and decom- 
posing; they are readily soluble in most of the ordinary organic 
solvents, and dissolve comparatively easily even in boiling light pet- 
roleum, separating again, on cooling, in lustrous, transparent prisms. 

The hydrazone is very unstable, and soon decomposes on exposure 
to light and air ; its behaviour towards hydrochloric acid seems to be 
similar to that of the parabromo-derivative described below. 

Digitized by VjOOQIC 


a'KeU>Utrdhydronaphihalene Para^amophenylhf/drazone, 

The preparation of this substance from the purified ketone and 
parabromophenylhydrazine is carried out exactly as described in the 
case of the preceding componndi interaction taking place very readily ; 
the almost colourless, oily product, which is precipitated on the addi- 
tion of water, crystallises immediately when treated with a little 
methylic alcohol, and is easily purified with the aid of this solvent, 
from which it separates, on cooling, in long, colourless prisms, or in 
massive, transparent crystals. 

It melts ^t 117 — 118^, when heated fairly quickly from about 
110% effervescing and decomposing, and it is readily soluble in cold 
ether, ethylic acetate, acetic acid, and benzene ; it also dissolves in 
boiling light petroleum, but, on cooling, separates again almost com- 
pletely in nodular aggregates of needles. It is more stable than the 
hydrazone, and does not change colour when exposed to light and air 
during several days. 

Attempts to regenerate the ketone by distilling the parabrom- 
hydrazone with moderately concentrated hydrochloric acid were not 
successful ; under these conditions, the bromhydrazone is converted 
into a crystalline compound, which is probably produced by a change 
analogous to that which occurs in the formation of " benzyleneindol " 
from the phenylhydrazone of a-hydrindone (compare B!ausmann^ Ber,^ 
1889, 22» 2019 ; Kipping, Trans., 1894, 65, 494). 

a-Ketoieirahydranaphihaleneoximet C^qH^qINOH. 

On gently heating a solution of the purified ketone in dilute 
methylic alcohol with hydroxylamine hydrochloride and excess of 
potash,, the separation of a crystalline compound soon commences, if 
the alcohol be sufficiently dilute, and on allowing the solution to 
evaporate on the water-bath, most of the product is deposited in 
colourless plates. It is easily purified by recrystallisation from dilute 
methylic alcohol, the lustrous, transparent, well-defined rhomboidal 
crystals thus obtained generally exceeding 10 mm. in diameter. It 
melts at 102*5 — 103*5° without decomposing, and is very readily 
soluble in cold ether, chloroform, mothylic alcohol, and most other 
solvents ; it also dissolves freely in cold concentrated potash, but it is 
insoluble, or nearly so, in water* 

Digitized by VjOOQIC 


Conversion qf a-Ketotetrahydronaphthalene into Tetrahydro^- 

Although the method of formation of the ketone descrihed above 
and the analysis of its semicarbazone left little room to doubt that 
it had the constitution assigned to it, we thought it would be inter- 
esting to try and convert it into one of the tetrahydronaphthalene 
derivatives of known constitution ; for this purpose, experiments were 
made with the object of converting the ozime into the tetrahydro-a- 
naphthylamine described by Bamberger and Bammann {Ber,^ 1889, 
22, 951). 

Attempts to reduce the ozime with sodium and moist ether were not 
successful j even after employing a large excess of sodium, a portion 
of the ethereal solution gave, on evaporation, crystals of the un- 
changed oxime, and a basic substance appeared not to have been 
formed ; we, therefore, tried the action of sodium amalgam in warm 
dilute acetic acid solution, and found that reduction took place very 
readily. After using the amalgam in considerable excess of the 
theoretical quantity, the acid solution was submitted to distillation 
with steam, but as no unchanged oxime passed over, the solution was 
rendered strongly alkaline with potash, and again submitted to steam 
distillation. A colourless, strongly basic oil, which was only 
moderately soluble in water, then collected in the receiver. This 
aqueous distillate, having been mixed with excess of hydrochloric acid 
and evaporated almost to dryness on the water-bath, left a consider- 
able quantity of a salt which crystallised in needles or prisms, and 
was very readily soluble in water ; the yield seemed to be practically 

In order to prove that this salt was the hydrochloride of tetra- 
hydro-a-naphthylamine, a portion of it was roughly dried at 100^, 
and then heated for a few minutes with excess of acetic anhydride ; 
on subsequently cooling and adding water, the acetyl derivative of the 
base separated in colourless needles. 

Bamberger and Bammann {loc, dt,) describe this acetyl derivative 
as crystallising from very dilute alcohol in felted masses of needles, 
and they give 148 — 149^ as its melting point. The compound we 
obtained had these and all other properties mentioned by Bamberger 
and Bammann, except that it melted at 144 — 145% and its melting 
point underwent no change on recrystallisation. As, therefore, there 
was this considerable difference in melting point, which is difficult to 
account for, further evidence as to the identity of our base was 

For this reason, we prepared the jp^o^inoeMorufo, a salt which is 

Digitized by VjOOQIC 


immediately precipitated in yellow plates or prisms on the addition of 
platinic chloride to a solution of the hydrochloride of the hase. This 
Gomponnd crystallised from water in long, flat, orange-yellow prisms, 
was readily soluhle in hot water and cold methylio alcohol, and melted 
immediately at 140° when suddenly heated, thus indicating that it 
contained water of crystallisation ; these properties agreed with those 
assigned to the platinochloride of tetrahydro-a-naphthylamine hy 
Bamberger and Bammann, and our salt, like theirs, melted at 190°, 
decomposing and effervescing; the temperature at which the salt 
melts and decomposes varies, however, within very wide limits, 
according to the rate of heating, and on heating quickly the tempera- 
ture can be raised to about 197° before liquefaction and decomposition 

An analysis was, therefore, necessary to confirm the supposed 
identity of the two compounds. 

0*4543 salt, dried over sulphuric acid, lost 0*0232 at 100°. HgO -= 51 . 

0*4311 anhydrous salt gave 0*1224 platinum. Pt » 28*4. 

Calculated for (CioHi3N)2,H2PtCIe + 2HjO, H20 = 4-87 per cent. 

Calculated for (CioHi3N)2,HjPtCle, Pt = 27*6 per cent. 

On ignition, the salt gave off fumes having a strong odour of 

Univkbsitt College, 

XVIII. — Production of Optically Active Mono- and 
Di-alhyloocysuccinic Adds from Malic and Tartaric 

By Thomas Pubdie, F.R.S., and William Pitkeathly, B.Sc., 
Berry Scholar in Science. 

It has been shown in previous papers (Trans., 1898, 73, 287, 862) 
that the high optical activity of the ethereal malates and lactates, 
prepared from (lie respective silver salts, is due to the production of 
small quantities of the more active alkylozy-derivatives, and Bodger 
aQd Brame (Trans., 1898, 73, 306) were consequently inclined to 
attribute the abnormally high rotation of the ethereal tartrates, 
which they prepared from silver tartrate, to a similar cause. On 
hydrolysing specimens of methylic tartrate, on the other hand, which 
had been prepared by this and by the commoner methods, they found 
that the products diowed practically the same rotation, which sug- 

„.gitized by Google 


gested the possibility that the ethereal salts from the two different 
sources might be isomeric. 

The object of the present research was to obtain further evidence 
of the production of alkylozy-derivatives in the interaction of the 
silver salts of hydrozy-acids with alkyl iodides, and to find a modifi- 
cation of the reaction which might serve as a practical method for their 
preparation. We have, accordingly, made further experiments on 
the production of alkyloxysuccinates from silver malate, and, Mr. 
Brame having kindly left us to continue his work, we have examined 
the product of the action of isopropylic iodide on silver tartrate for 
di-isopropoxysuccinic acid. We find that this compound is actually 
produced, and conclude, therefore, that a similar reaction oocurs, 
although probably to a smaller extent, when other alkyl iodides are 
used. We find further that the ethereal malates and tartrates can be 
readily alkylated by means of alkyl iodide in the presence of silver 
.oxide, the reaction furnishing a convenient method of preparing the 
optically active mono- and di-alkyloxysuccinio acids from malic and 
tartaric acid respectively. The ethereal di-alkyloxysuccinates are 
highly active compounds, and their presence even in small quantity 
would account for the abnormally high rotation of the ethereal 
tartrates prepared from silver tartrate. Observations made on the 
metallic salts of diethoxysuccinic acid also explain the apparently 
anomalous results obtained by Bodger and Brame in the hydrolysis of 
methylic tartrate. 

Action of Alkyl Iodides on Silver Malate, 

In former experiments with ethylic iodide {loe. eit.)f the crude 
distilled product of the reaction had always nearly the same rotation, 
-U° (^=1). We now find that, by modifying the conditions, a 
somewhat more active liquid is obtained, although in no case is the 
proportion of ethoxysuccinate produced large enough to admit of its 
being separated from the malate. Thus, by adding the iodide (3 mols.) 
gradually to the silver malate (1 mol.), the treatment being otherwise 
the same as before, a product was obtained which, without fractional 
distillation, gave the rotation — 15-31°. Even when 2 mols. of iodide 
only were used and benzene added to moderate the action, con? 
ditions which should be unfavourable to the hydroxyl group being 
attacked, the ethoxysuccinate was formed as before, the ethereal salt 
obtained showing a rotation almost identical with that just quoted. 

In attempting to prepare isobutylic malate by the silver salt 
method (Trans., 1896, 69, 824), the product was found to consist 
mainly of free acid, which, however, was not further examined at the 
time; as it seemed possible that this result might be due to the 

Digitized by VjOOQIC 


formation of isobutoxysuccinic acid, we have investigated the reaction 
more closely. The method employed was the same as before, but the 
.free organic acid was removed from the product before distillation 
by allowing it to stand over potassium carbonate. In an experiment 
in which 92 grams of malate were added gradually to 167 grams of 
iodide, the yield of ethereal salt amounted to only 6 grams, and in 
another experiment, in which the order of mixture was reversed, the 
yield was still less. The free organic acid, which was isolated as in a 
previous similar case (Trans., 1898, 73, 299), gave, on heating with 
a solution of barium hydroxide, a large quantity of barium malate ; 
when dried at 160^, it was found to contain 50*89 per cent, of barium, 
the calculated number being 50*93. The filtrate from this, on 
evaporation, left a less granular and more soluble salt, which analysis 
showed to be barium isobutoxysucoinate contaminated with some 
malate. Great difficulty was experienced in freeing the isobutoxy- 
suocinic add from malic acid. We found, finally, that although both 
barium salts are precipitated on boiling their aqueous solutions, the 
isobtttoxysnocinate differs from the malate in redissolving readily 
when the solution cools, which enabled us to obtain the isobutoxy- 
suodnate as a flocculent powder approximately free from malate. The 
results of the combustion of the substance, dried at 160^, were as 

I. C = 28-93. H = 4*08 ; Ba = 42*71 per cent. 
11.0 = 29*03. H- 405 ; Ba = 42*36 „ „ 
CgH^OjBa requires = 29*64 ; H = 3*69 ; Ba = 4216 per cent. 

In very, dilute aqueous solutions (e = 0*9132) at 13°, the saH 
showed the specific rotation -21*4°. The silver salt could not be 
obtained by precipitation. The Imodium salt, prepared from the barium 
salt, gave in aqueous solution the specific rotation - 27*8° (e= 1*888), 
and, on analysis, was found to contain 19*46 per cent, of sodium, instead 
qf the calculated number, 19*66. The rotations quoted are much 
higher than those of the corresponding nuilates, and are such as 
butoxysuocinates might be expected to exhibit. 
. The 5 grams of ethereal salt, mentioned above, contained a small 
proportion of isobutylic isobatoxysuccinate. This was evident from 
its high observed rotation, - 16*28° in a 100 mm. tube, that of the 
oorresponcUng malate being only - 11*60° {ZeU. physikal. Chem., 1895), 
17, 249), and was confirmed by the detection of barium isobutoxy- 
suodnate in the product of hydrolysis. The quantity of substance 
was too small to admit of its being purified, but it showed a specific 
rotation nearly the same as that quoted above, namely, - 20*81° at 
10° for cs^ 1*638. These observations show that the reaction between 
isobutylic iodide and silver malate does not follow the normal course, 

Digitized by VjOOQIC 


but that the main product is free malic acid, produced probably by 
the decomposition of the iodide into isobutylene and haloid acid, and 
that a considerable quantity of isobutoxysuccinio acid is also formed. 
Mr. J. 0. Irvine was good enough to examine for us the action of 
secondary butylic iodide on silver malate, in the hope that this iodidci 
like isopropylic iodide, would yield a larger proportion of the alkyloxy- 
acid. He obtained from the product only a small quantity of a very 
soluble barium salt of wax-like appearance which, although it could 
not be obtained quite pure, was evidently a barium butoxysuccinate. 
On combustion, it gave results approximating to the calculated 
numbers, and its specific rotation in dilute aqueous solution, - 20*97% 
was practically the same as that of the isobutoxysuccinate. 

Action of liopropylic Iodide on Silver Tartiraie, 

Isopropylic iodide was used in this experiment, previous results having 
shown that it is more prone than other iodides to the reaction by which 
the alkyloxy-acid is produced. Sixty-eeven grams of silver tartrate 
(1 mol.) were gradually added to 134 grams of the iodide (about 4 mols.) 
previously diluted with an equal volume of benzene, and the mixture was 
treated as in previous similar experiments. After distilling off the bens- 
ene and unaltered iodide, and shaking the residual liquid with a solution 
of sodium carbonate, a small quantity of an oily, ethereal salt remained, 
which, as it was too dark coloured for polarimetrio observation, was 
diluted with an equal bulk of alcohol and then examined in a 100 mm. 
tube. The observed rotation, about + 26% showed that the substance 
was far more active than any of the ethereal tartrates. A crude, 
syrupy acid, obtained from this by hydrolysis with potassium hydroxide, 
acidifying with sulphuric acid, and extracting with ether, showed the 
specific rotation + 29^ in about a 10 per cent, solution, an activity 
much exceeding that of tartaric acid. The barium and magnesium salts 
did not crystallise ; the calcium salt was very soluble in cold water, 
but was precipitated as a crystalline powder on boiling its aqueous 
solution. Estimations of calcium in the salt, dried at 130% gave the 
results 14*77 and 14*82 per cent., the calculated percentage for 
calcium di-isopropoxysuccinate being 14*70. 

The production of the alkyloxy-compound in the reactions, which 
have been described in this and previous papers, appears to be due to 
some of the ethereal salt of the hydroxy-acid, which is formed in the 
first instance, reacting further with alkyUc iodide and unaltered silver 
salt with the formation of silver iodide, ethereal salt of the alkyloxy- 
acid, and free hydroxy-add, which is always present in the product. 
The free alkyloxy-acid or its acid alkyl salt, which also frequently 
results from the reaction, may be produced from the normal alkyl 

Digitized by VjOOQIC 


salt by water accidentally present, or by a reaction such as the 
following, in the case, for example, of silver malate and ethylic iodide, 
OH- CgH3(C00Ag)j + 2EtI = OEt-0jH5(0OOEt)*0OOH + 2AgL 

Consideration of the reaction in qnestion suggested that the alkyla- 
tion would probably be much more complete if the ethereal salt of the 
hydrozy-add was first prepared and then treated with alkyl iodide in 
the presence of silver oxide, or other metallic oxide of a similar nature. 
Ethylic malate, ethylic iodide, and lead oxide gave a n^ative result, 
but the product obtained by heating the malate and iodide with 
mercuric oxide yielded, on distillation in a vacuum, an ethereal salt 
much more active than the original malate, the rotation having risen 
from -11-8° to —30*7° (^"=1)- A mixture of ethylic malate and 
isopropylic iodide reacted very vigorously with silver oxide, and gave 
an ethereal salt showing the rotation - 30*5°. As these results indi- 
cated that the expected reaction had occurred, the action of ethylic 
iodide on ethylic malate and on ethylic tartrate in presence of silver 
oxide was examined in detaiL 

Preparcttion of BthyUe EihoxyiuccvnaU from Ethylic Malate. 

The materials used were 52 grams of ethylic malate, the rotation 
of which was a« - 11-83° (?=1, < = 6°), 86 grams of ethylic iodide, 
and 64 grams of dry silver oxide, these proportions being chosen on 
the assumption that the reaction occurs in accordance with the 
equation OH'C3H3(OOOEt)j + 2EtI + Ag30= OEt-02Hj(COOEt)2 + 
EtOH + 2AgI. The mixture, which underwent an energetic reaction 
on being gently warmed, was afterwards heated for some time on a 
water-bath, then diluted with benzene, filtered,* and distilled under 
reduced pressure, when it yielded 43 grams of a liquid having a 
nearly constant boiling point, and showing the rotation— 41*21° in a 
100 mm. tube at 6°. Assuming that the liquid consisted of only 
ethylic ethoxysuccinate and malate, the rotation indicated the 
IHreeence of about 60 per cent, of the former compound. To remove 
the malate, the mixture was shaken repeatedly with a cold 10 per 
cent, aqueous solution of potassium hydroxide until the residual oil 
was reduced to about one-half its original weight. The oil, after 
being washed with water and dried with calcium chloride, was found 
on distillation to boil at the same temperature as ethylic c?-ethoxysuc- 
cinate formerly prepared (Trans., 1895, 67, 972), namely, at 124° 
under a pressure of 10 mm., and its analysis gave the following 
results agreeing with the composition of that substance. 

* The silyer residues from these and the succeeding experiments were black and 
their composition remains to be examined. 


Digitized by VjOOQIC 


Found : I. 0«64'63 ; H = 8-68 per cent, 
n. = 64-79; H = 8U „ „ 

Calculated : Oa66'06; H»8*26 per cent. 

A determination of the specific rotation of the liquid at 6^ gave the 
following result: o= -66-85°, l^l, d 6^4^ « 1-0601, hence [aj^- 
" 64*14^. The specific rotation of ethylic c^thozysucoinate from the 
active acid, which was obtained by resolution of the raoemoid com- 
pound {loe. eU.9 p. 979), was +66-48^. The somewhat lower rotation 
now found is accounted for by the substance being contaminated 
with some ethylic f umarate, produced probably in the preparation of 
the ethylic malate. The specific rotation of an aqueous solution of 
l-ethoxysuecime cicid, which was obtained by hydrolysing the ethylic 
salt, acidifying the product and extracting it with etiier, was found to 
be [u]d= -31-14° (c = 80588, < = 7°), a number about 3° lower than 
that previously found, the discrepancy here being greater than in the 
case of the ethylic salts owing to the fumaric acid being chiefly con- 
tained in the crystallised portion of the acid which was used in the 
determination. For further comparison of the acids from the two 
sources, observations were made on the otcid ammorUum saU which 
could be freed by crystallisation from f umarate. A determination of 
the specific rotation of the hydrated salt in aqueous solution at 8° gave 
the foUowing result: a= -6-00°, /=2, c = 10'0288, hence [a]©- 
— 29-91°. The value formerly found for similar concentration at 17° 
was 28*46°. To obtain evidence of the purity of the compound, a 
silver salt was made from it, and analysed with the following result 

Found: = 18-90; H = 2-27; Ag-67'76 and 67-69 per cent. 
OeHgOjAgj requires 0«19-16 ; H«2-13 ; Ag = 67-46 per cent. 

The yield of alkyloxysuccinic acid by the process which has been 
described would probably be increased by employing a larger propor- 
tion of alkyl iodide and silver oxide, in order to allow for the loss which 
is doubtless entailed by their partial direct interaction. 

Preparaiicn of Ethylic di-Diethoxy succinate from Ethylic Tartrate. 

Attempts have been frequently made to alkylate the alcoholic 
hydroxyl groups of tartaric acid, but without success. Perkin (Trans., 
1867, 20, 166), by the action of ethylic iodide on ethylic monosodio- 
tartrate, obtained an oil which he thought was probably ethylic mon- 
ethyltartrate, but the later researches of Lassar Oohn {Ber,^ 1887, 20, 
2003), Mulder {Rec. Trav. Chim., 1889, 8, 361) and Freundler (BuU. 
Soc. Chim.y 1894, 11, 308) have shown that neither the sodium nor 
potassium derivatives of ethylic tartrate react in the usual way with 
alkyl iodides. The ethylic diethoxysuccinate which Michael and 

Digitized by VjOOQIC 


Booher {Ber.y 1896, 29, 1792) found to be one of the products of the 
action of sodium ethoxide on ethylic dibromosuccinate and on ethylic 
acetylenedicarboxylate, was shown by these authors to be the unsym- 
metricai compound. 

The proportions of ethylic iodide and silver oxide employed in our 
experiments were 6 mols. of the iodide and 3 mols. of the oxide to 
1 mol. of the tartrate, that is to say, an excess of one-half over the 
calculated quantity, assuming the reaction to proceed in the sense 
indicated in the case of ethylic malate. On adding the oxide to the 
mixture of iodide with tartrate, the reaction set in spontaneously, 
and became so violent that it had to be moderated by cooling. The 
liquid product, obtained as described under ethylic ethoxysuccinate, 
was about equal in quantity to the tartrate employed, its boiling 
point was nearly constant, and its observed rotation varied in different 
preparations from +83*5^ to +92*5^ in a 100 mm. tube. As the 
percentage numbers found, on analysis, were somewhat lower than 
those calculated for ethylic diethoxysuccinate, and as it proved 
impossible to purify the substance either by fractional distillation or 
by shaking with water, which we expected would remove unaltered 
ethylic tartrate, if present, we had to resort to the method of partial 
hydrolysis by shaking with a 10 per cent, aqueous solution of 
potassium hydroxide. The process entailed a considerable loss of 
material, but the residual oil, on redistillation (b. p. 149 — 151° at 
15 mm.), had increased in rotation to 4-97*5°, and on analysis gave 
results in agreement with the calculated numbers. 

Found : L = 54-99 ; H-8-71 per cent. 
IL 0» 64-87; H-8-67 „ „ 
Oalculated for CuH^jOg : « 5496 ; H =» 840 per cent. 

The specific rotation of the liquid at 18° was as follows : a « + 97-52°, 
/=1, d 18°/4°« 1-0460, hence [a]D= +93-23°. The ethereal salt of 
lower activity, which was removed by the partial hydrolysis, gave an 
uncrystallisable acid, which we intend to examine further. 

d'DidhoxtfSueeinie ctcid, which was obtained from the ethylic salt in 
the same manner as the monethoxy-acid, is sparingly soluble in 
benzene, readily soluble in ether, alcohol, chloroform, and water, 
from which it crystallises in long prisms melting at 126 — 128°. The 
results of the analysis of the substance, dried at 100°, were as follows. 

Found : I. = 46-86 ; H = 704 per cent. 
II. 0=46-81; H = 6-90 „ „ 
Oalculated for C^H^fi^ : = 46-60 ; H = 6-80 per cent. 

The acid showed the following specific rotations in aqueous solu- 

M 2 

Digitized by VjOOQIC 


tions at 20°: a- +6-73, c«10-U88, Ul, hence [a]i»=« +66-31°; 
a- +2-70°, c-4-0596, ^=1, hence [a]D= +66-51° 

The silver salt is soluble in water, and does not decompose much on 
evaporating the solution. An estimation of silver in the salt gave 
the result 51-01 percent., the calculated number being 51-43. The 
acid potassium and acid ammonvwn salts are crystalline. The normal 
sodium salty in aqueous solution at 17° showed the specific rotation 
+ 41*11° (cs 3*138) j the residue left on evaporating the solution, 
dried at 120°, was found to contain 18-35 per cent, of sodium instead 
of the calculated number^ 18*40. The ealeium salt is very soluble, and 
was not obtained in the crystalline state. The barium salt, which is 
characteristic, is sparingly soluble in cold water, and crystallises 
readily in large, glassy prisms, containing apparently 4HsO, which is 
lost at 120—130°. Analysis of the anhydrous salt gave the following 

Found: Oe 27*78; H»3-77; Ba« 39*94 and 4012 per cent. 
Oalouhited: 0-28-15; H«3-52; Ba-=40*18 „ „ 

An aqueous solution of the salt at 16° had the specific rotation 
+ 26 25° (c= 1*8092). 

Ethylic (i-diethozysuccinate is also obtained by the direct action 
of ethylic iodide and silver oidde on tartaric acid, but the yield is 
smaller than when the alkyl tartrate is used as the starting point. 
From 17 grams of tartaric acid we obtained in this way 12 grams of 
an ethereal oil having the same boiling point, and the same observed 
rotation, + 85°, as the crude ethylic diethozysuccinate prepared from 
ethylic tartrate. 

The optical effect of the replacement of the alcoholic hydrogen of 
tartaric acid by alkyl radicles is of the same nature as that which 
attends a similar substitution in lactic and malic acids ; the sign of 
rotation, defined in the case of lactic acid as that of its salts, remains 
unchanged ; a striking rise of activity, observable more particularly 
in the free acids and the ethereal salts, is produced, and the specific 
rotation of the acids in aqueous solutions of varying concentration 
becomes more constant. Thus, in passing from tartaric to diethozy* 
succinic acid the molecular rotation of the ethylic salt is raised from 
+ 15*8° to + 244-3°, and that of the free acid at similar concentration 
from +20*6° to +136*6°. The ionic rotation, however, does not ez^ 
perience a proportional rise, the result being that, whilst in the case 
of the three hydrozy-acids mentioned the molecular rotations of the 
alkali salts in dilute solution greatly ezceed those of the free acids, 
these rotations become nearly equal in the case of the alkylozypro^ 
pionic compounds, and the order of their value is rever^ in ^he cetse 
of the mono- and di-alkylozysuccinic compounds. 

Digitized by VjOOQIC 


Our observations furnish a satisfactory explanation of the apparently 
anomalous results obtained by Rodger and Brame, which have been 
already referred to. An admixture of 6 per cent, of ethylic cMi 
ethoxysQocinate would suffice to account for the high rotation of the 
ethylic tartrate which they prepared by the silver salt method, but 
the difference between the rotations of the products of hydrolysis of 
such a mixture and of pure ethylic tartrate respectively would not, 
under the conditions of the experiment described by them, amount to 
more than O'l— 0-2®. • 

The action of alkyl iodides and silver oxide on the alkyl salts of 
optically active hydroxy-acids furnishes a general method of obtaining 
the active alkyloxy-acids. We have used the method with success for 
the preparation of alkyloxypropionic and alkyloxyphenylacetic acids 
from active lactic and mandelic acids, and we are at present studying 
the application of the alkylating agent to the alkyl tartrates in 
general and to other compounds. 

Ukitsd Collbob of St. Salvator and St. Lbonard, 
Univbbsity of St. Andrews. 

XIX. — Determination of the Constitution of Fatty 

Acids. Part L 

By Arthur William Cbosslby and Hsnrt Bondbl Lb Sueur. 

Some short time ago one of us, in conjunction with Professor Ferkin 
(Trana, 1898, 73, 1), described an investigation of a complicated 
mixture of fatty acids derived from the fusion of camphoric acid 
with alkalis ; as the difficulties encountered in identifying some of 
the fatty acids were very great, it was considered desirable to try 
and devise a method for the determination of the constitution of 
such acids, and the object of this paper is to give a short account of 
experiments which have been carried out in this direction. 

So far as our present experiments go, we think they may be 
described as satisfactory, and although the method may not be an 
infallible one, it seems likely to prove of considerable importance as 
a means of establishing the constitution of organic acids. 

The method of procedure which was suggested to us by Professor 
Perkin is the foUowing. Starting with a fatty acid, X*CK^* CH,* COOH, 
this is first converted into the ethylic salt of the monobromo-deriva- 
tive by Yolhard's process (^ntio/an, 1887, 242, 61) ; from the work of 

* The experiment referred to was made on the methylic tartrates, bnt this wonld 
not materially alter the result as stated above. 

Digitized by VjOOQIC 


Auwers and Bemhardi {B$r,f 1891, 24, 2209) and others, there can 
be no doubt that, under these conditions, the bromine atom takes up 
the a-position, yielding the substance X- CH,* ^^^^'^^^^^^^^* ^^^* 
brom-ethereal salt is then treated with quinoline or diethylaniline 
(compare Weinig, Annalm^ 1894, 280, 253), whereby the elements 
of hydrobromic acid are removed, and the ethylic salt of an unsatu- 
rated acid of the acrylic series, X*OHIGH*COO£t, is produced. The 
free acid obtained from this salt by hydrolysis is then oiddised, first 
with potassium permanganate, giving rise to the corresponding 
dihydroxyacid, X-CH(OH)*CH(OH)-COOH, and then with chromic 
acid, when the molecule is broken down at the position occupied by 
the double bond in the unsaturated acid, giving X*COOH and 

The result is, therefore, the production of oxalic acid and a fatty 
acid (or ketone) containing two carbon atoms less than the original 
acid, and as the number of isomerides decreases greatly with loss of 
two carbon atoms, the possibility of identification is much enhanced. 
We have carried out this process with three acids, namely, valeric, 
t9o-valeric, and wo-butylacetic acids, as we considered that the 
oxidation products of the acids of the acrylic series corresponding 
with these fatty acids would be typical examples of what might be 
expected to be met with in actual determinations. Thus valeric acid 
gives ethylacrylic acid, and this, on oxidation, propionic acid {fwrmal 
acid) ] Mo-valeric acid gives dimethylacrylic acid, which then yields 
acetone (hUone) ; and wo-butylacetic acid gives wo-propylacrylic acid, 
and, on oxidation, ifo-butyric acid (UfHioid), 

With the three acids mentioned, the method works well. In all 
cases we have been careful to note the yields of substances obtained 
in the various stages, and there is h^re appended a tabulated list of 
results. The numbers express the percentage yields, and are referred, 
in the case of the unsaturated ethereal salt and acid, to the amounts 
theoretically obtainable from the brom-ethereal salt employed ; and 
in the case of the " acid or ketone produced on oxidation," to the 
amounts theoretically obtainable from the unsaturated acid used. 

Acid employed. 






Acid or ketone 

produced on 


Valeric acid 




66— (M) 



iso-Valeric acid 

iso-Batylacetic acid. 


The poor yield of unsaturated ethereal salt obtained from ethylio 
bromovalerate is accounted for by the fact that thete is also pro- 
Digitized by VjOOQIC 


ducod a coiisi4erabIe quantity of some substance of higher boiling 
point, which is at present under investigation ; and the comparatively 
small amount of the solid dimethylacrylic acid obtained from its 
ethylic salt is due to the fact that an oily substance is produced at 
the same time (see page 164). In no case was oxalic acid identified 
in the products of the reaction, nor could it be expected to resist the 
action of the strong oxidising agents employed. 

We have experimented with both quinoline and diethylaniline, 
using them as reagents for the elimination of the elements of hydro- 
bromic acid, and £nd that, with the lower fatty acids, the one gives 
quite as good results as the other, but with the higher fatty acids 
diethylaniline is to be preferred. For example, in the case of ethylic 
bromisobutylacetate, the 42 per cent, yield of ethylic isopropyl 
acrylate, obtained when using quinoline, was increased to 70 per cent. 
by employing diethylaniline. Quinoline always gives rise to tarry 
products, which are not easy or agreeable to work with, whereas 
diethylaniline does not ; in the latter case, however, the substances 
require to be heated together for a much longer time, and it is very 
difficult to eliminate the last traces of hydrobromic acid. As, how- 
ever, the unsaturated ethereal salts are subsequently heated with 
alcoholic potash, the latter objection is of no great moment. 

We intend to further test the efficacy of the method by trying it on 
other acids, such as (a) stearic acid, and from preliminary experi- 
ments already made with this acid, it seems highly probable that the 
various reactions will take place as expected. The insolubility of the 
hydroxylated higher fatty acids in water may render the oxidation 
with chromic acid a difficult operation, in which case it will be of 
interest to see whether fusion with potash will serve a similar purpose. 

(b) It will be noticed that, among the acids examined, none contain 
BXkyl groups in the a-position. In such a case, as, for example, ethyl- 
isopropylacrylic acid, quite a new point is raised. This acid, 

still contains one a-hydrogen atom, and should, therefore, yield an 
a-brom-ethereal salt in the usual manner; but when the latter is 
trei^ted with quinoline, there are two possible ways in which hydro- 
bromic acid may be eliminated (compare Perkin, Trans., 1896, 09, 
1466), giving rise to 

(CH8)2CH-(:|-COOH (CH3)2C:(p- COOH 



Methyl-a-iaopropylacrylic acid. Dimethyl-a-ethylacrylic acid. 

The study of the oxidation products of the unsaturated acid or 
acids produced would, therefore, be of special importance. 

Digitized by VjOOQIC 


(c) The method may also prove to be applicable in the case of di- 
basic acids, and we propose to try it on pimelic (isopropylsuccinic) acid. 


Acetone from Isovaleric Acid. 

Instead of starting with isovaleric acid, the ethylic salt of a-brom- 
isovalerate supplied by Kahlbaum was employed,yhich, on distillation, 
boiled constantly at 185 — 186^, and a bromine determination gave 
the following numbers. 

0-2514 gave 02275 AgBr. Br = 38-50. 

C^HgBr-COOOgHg requires Br = 38-28 per cent. 

Treaiment of Ethylic a-BromiaovcUerate toith QuinoUne, — Weinig (An- 
nalen, 1894,280, 253) has shown that diethylaniline may be used instead 
of alcoholic potash for the elimination of the elements of hydrobromic 
acid, and later Perkin and Goodwin (this Journal, 1896, 69, 1470) 
described experiments in which they employed quinoline for the 
preparation of dimethylacrylic acid from ethylic a-bromisovalerate. 
We have followed their instructions exactly, using 50 grams (1 mol.) 
of the brom-ethereal salt and 70 grams (2 mols.) of freshly distilled 
coal-tar quinoline ; on fractionating the product, it was found to distil 
between 153^ and 155^ as a colourless oil of penetrating odour. 
The yield is 80 per cent, of the theoretical. 

This ethereal salt is readily saponified by alcoholic potash, and the 
dimethylfiorylic acid formed boils constantly at a temperature of 114^ 
under a pressure of 40 mm. On standing, the distillate solidifies 
almost completely to a mass of needle-shaped crystals, which, after 
being freed from the mother liquor by spreading on a porous plate 
and recrystallisation from light petroleum (b. p. 60 — 80^), melted at 
68*5 — 69^, and gave the following results on analysis. 

0-1047 gave 02296 CO, and 0-0744 H,0. C = 5980 ; H = 7*90. 
CfiHgOg requires 0= 60-00 ; H«8-00 per cent. 

The yield of pure acid is from 55 — 60 per cent, of the theoretical 
obtainable from the brom-othereal salt employed. On extracting the 
porous plate just mentioned with ether, a small amount of an oily 
liquid was obtained which showed no signs of solidifying even after 
long standing, and which was not further investigated. 

Perkin and Goodwin {loc, cit,) also mention this oily bye-product. 

Treatment of Ethylic a-Bromieovalerate with Diethylaniline. — In using 
quinoline for the elimination of the elements of hydrobromic acid, 
there is always a considerable quantity of tarry matter formed, and 
in later experiments we found that diethylaniline could be used with 



advantage instead of quinoline; the yield of unsaturated ethereal 
salt is considerably increased, and no tarry products are produced, 
although the mixture requires to be heated for a much longer period. 
The yield of ethylic dimethylacrylate as obtained in the experiments 
just described is good, but we thought it of sufBicient interest to try 
the effect of diethylaniline on ethylic a-bromisovalerate. The process 
was carried out exactly as described on page 166. 

There is, however, in this case, no particular advantage to be de- 
rived, for the yield of ethylic dimethylacrylate is no higher than when 
quinoline is employed. 

Oxidation qf JHmethylacryUo Acid with Potiunum Permangcmate, — 
Ten grams of dimethylacrylic acid were neutralised with potassium 
hydroxide and dissolved in 600 c.c. of water ; the whole was stirred 
with a turbine, and maintained at 0^ throughout the operation ; a rapid 
current of carbonic anhydride was then passed in, and a cold solution 
of 12 grams of potassium permanganate in 400 c.c. of water gradually 
added from a tap funnel. After standing overnight, the liquid was 
filtered from precipitated manganese dioxide, and evaporated to a 
small bulk; no attempt was made to isolate the dihydroxy-acid 
produced, but the liquid was at once submitted to the process of : — 

Oxidation with Potassium Diehrovnate and Svlphwiic Aeid. — For this 
purpose, the evaporated liquid was placed in a flask connected with a 
condenser, and, after warming to 70^ on the water-bath, a solution of 
30 grams of potassium dichromate in dilute sulphuric acid was slowly 
run in, the heating continued for eight hours, and the whole steam 
distiUed. The distillate, which was slightly acid to litmus paper, was 
carefully neutralised with potassium hydroxide, and again steam dis- 
tilled ; on adding an alcoholic solution of parabromophenylhydrazine 
to the distillate, there was an immediate and copious precipitate, 
which was collected and dried on a porous plate. After recrystallisation 
from light petroleum (b. p. 80 — 100^), it was obtained as beautiful, 
shining scales melting at 94^ Bromine determinations gave the 
f oUowing results. 

L 0-2886 gave 0*2314 AgBr. Br » 3409. 
II. 0-1226 „ 0-0998 AgBr. Br » 34*58. 

(OH3)sO:N*NH-O0H4Br requires Br » 35*24 per cent. 

Although the figures obtained are not so good as might be desired, 
there can ^be no doubt that this substance is the parabromophenyl- 
hydrazine compound jof acetone (compare Neufeld, AnnaUn, 1888, 248, 
96), the ease with which it undergoes decomposition accounting for 
the lowness of bromine found. 

The parabromophenylhydrazine compound obtained weighed 10*7 
grams, whereas the amount theoretically obtainable from 10 grtims of 

„.gitized by Google 


dimethylacrylic acid is 22*7 grams ; this is a 47 per cent, yield of the 

The residue from the steam distillation was examined for oxalic 
acid, after removing the chromium by the use of sulphurous acid and 
then boiling with excess of sodium carbonate. No traces of the acid 
were found, nor is this to be wondered at, as, when produced, it would 
at once be further oxidised to carbonic anhydride and water, in 
presence of the strong oxidising agent employed. This remark also 
applies to the other oxidations mentioned in this paper ; in no case 
was any oxalic acid detected. 

Propionic Add from Valeric Acid. 

Preparation qf Valeric Acid, — This acid was prepared by the con- 
densation of ethylic sodiomalonate and normal propylic iodide, and, 
after saponification, heating the propylmalonic acid so formed. It 
boiled constantly between 184° and 185° (uncorr.). 

Brominatian qf Valeric Add, — Forty-four grams of valeric acid were 
brominated with 135 grams of dry bromine and 4*5 grams of amorphous 
phosphorus in the usual manner, the product slowly poured into three 
times its volume of absolute alcohol, and the ethylic a-bromovalerate 
extracted with ether and distilled under diminished pressure. It 
boils constantly at a temperature of 110° (40 mm.), and is obtained 
in nearly theoretical amount (97*5 per cent.). 

I^reaiment qf Ethylic a-Bromovalerate tvith DiethylanUine. — Seventy- 
six grams (1 mol.) of the brom-ethereal salt, and 115 grams (2 
mols.) of diethylaniline, were heated in two portions in a flask 
attached to an air condenser and containing a thermometer, so that 
the temperature of reaction could be noted ; this begins at about 175°, 
and the thermometer rises rapidly to 190 — 200° at which temperature 
the whole was maintained for 6 hours. The product, when cold, 
was poured into excess of dilute hydrochloric acid, and extracted with 
ether, <bc., but as it was found to contain bromine, it was again heated 
with diethylaniline (1 mol.) for 16 hours at 190—200°. After ex- 
tracting in the manner just described, the ethereal solution was dried 
over calcium chloride, the ether distilled off, and the liquid residue 
fractionated ; by this means, 24 grams of ethylic ethylacrylate were 
obtained boiling at 155—160°. This is only a 52 per cent, yield, 
which is accounted for by the fact that 12 grams of some substance 
of higher boiling point (270 — 280°) was abo produced, the nature of 
which is at present under investigation. 

Ethylic ethylacrylate is readily saponified by alcoholic potash, yield- 
ing eihylacrylic add as a colourless, oily liquid, with pungent, charac- 
teristic odour, boiling at 195 — 197° under atmospheric pressure, and 

Digitized by VjOOQIC 


showing no signs of solidification even when cooled to - 14^. The 
yield of pure acid is 41 per cent, of the amount theoretically obtain- 
able from the ethjlio bromovalerate employed. 

A portion of the add was converted into the silver salt and 
analysed. . 

0-3510 gave 01834 Ag. Ag» 52-25. 

C5H702Ag requires Ag = 52*17 per cent. 

Oxidation of EihyUwrylic Acid. — Fourteen grams of the acid were 
oxidised exactly as described on page 165, firstly, with 16 grams of 
potassium permanganate, and then with a solution of 45 grams of 
potassium dichxomate dissolved in dilute sulphuric acid ; the mixture 
was distilled in a current of steam, and the distillate, after neutralisation 
with potassium hydroxide, was evaporated to complete dryness, finally 
on the water-bath. The fatty acids obtained on distilling the residue 
with concentrated sulpburi^ acid were dried by standing for some 
time in contact with concentrated sulphuric acid, and then fractionally 
distilled. Eventually, nearly the whole of the acids boiled between 
137° and 143°, and a portion boiling at 140° was converted into the 
silver salt and analysed. 

0-1914, on ignition, gave 0-1 144 Ag. Ag = 5977. 

0-2386 gave 0-1744 CO^, 0-0640 H^O, and 01420 Ag. = 1992 ; 

H-2-97; Ag« 59-51. 
OjHjCOOAg requires C = 19-89 ; H « 276 j A.g = 5966 per cent. 

The fraction boiling between 137° and 143° was then converted into 
the anilide by heating for 24 hours with twice its volume of pure 
aniline, and the solid product was repeatedly crystallised from light 
petroleum (b. p. 80 — 100°), from which it separates in white, glisten- 
ing plates melting at 102 — 103° (compare Crossley and Perkin, Trans., 
1898, 73^ 34). 

0-1814 gave 15-2 c.c. moist nitrogen at 21° and 762 mm. N = 9*56. 
CjHg-OO-NH-CeHg requires N = 9-39 per cent. 

These data prove conclusively that the acid produced by oxidising 
ethylacrylic add m the manner described is propionic acid. The 
amount of propionic acid obtained was 5*5 grams, which corresponds 
with a 53 per cent, yield of the amount theoretically obtainable from 
the ethylacrylic acid used. 

I9chwbyrie Acid from laobtUylcioetic Acid. 

ProparotAon qf lactmfylaoetic Acid. — ^This acid was prepared by the 
condensation of ethylic sodiomalonate with isobutylic bromide, and 

Digitized by VjOOQIC 


subsequent distillatiou of the product, after saponification with alco- 
holic potash. The acid boiled at 198—200''. 

Braminatian of Ischutylacetic Acid, — Fifty-two grams of the acid, 
4 '8 grams of amorphous phosphorus, and 140 grams of dry bromine 
were treated in the usual manner. Ethylic a-iromisobuiylaeetate was 
obtained as a colourless, pleasant smelling liquid boiling at 115^ under 
a pressure of 43 muL 

01530 gave 0-1296 AgBr. Br = 36-01. 

CgHjoBr-OOOCgHg requires Br = 35-87 per cent. 

The yield of the ethylic salt was 93 per cent, of the theoretical. 

Treatment of EthyUc BromiaobutylcteeUUe with QuinoHne or Diethyl- 
aniline, — Eighty-five grams of the bromethylio salt were treated in 
two portions with quinoline exactly as described on page 164, and the 
unsaturated ethereal salt was distilled, when 22 grams were obtained 
boiling at 168 — 169° at the ordinary pressure. As this is only 42 per 
cent, of the amount theoretically obtainable, the effect of diethyl- 
aniline on the brom-ethereal salt was tried, when the yield was in- 
creased to 70 per cent. It was, however, found exceedingly difficult to 
get rid of the last traces of halogen ; and even after a third treat- 
ment with diethylaniline, the substance did not give good results on 
analysis. Possibly this is due to the presence of traces of halogen. 

I. 01040 gave 0-2626 C02and 0-0966 HgO. C-66-23; H = 10-31. 
n. 0-1670 „ 0-3820 CO2 „ 01416 HjO. C-66-36; H« 10-02. 
CfiH^-COOOgHfi requires C = 67-60 ; H = 9-86 per cent. 

Perkin and Goodwin (Trans., 1896, 69, 1471) found that ethylic 
dimethylacrylate, prepared in a similar manner, did not give good 
results on analysis, but no explanation of the fact is offered. 

When hydrolysed with alcoholic potash, the oily, unpleasant smell- 
ing liquid boiling at 169° is converted into ieopropylacrylie aoid, which 
is a colourless, oily liquid of exceedingly unpleasant odour, boiling at 
133° (60 mm.) ; it does not solidify when cooled to - 16° The yield 
is 61 per cent, of that theoretically obtainable. 

The silver ecUt is a white, amorphous, insoluble precipitate* On 
analysis, it gave the following numbers. 

0*2636 gave, on ignition, 0-1232 Ag. Ag «= 48-68. 

O^HgOjAg requires Ags=4g-77 per cent. 

Oxidation qf Ieopropylacrylie Acid, — ^Twenty grams of the acid were 
oxidised with 30 grams of potassium permanganate, and afterwards 
with 60 grams of potassium dichromate dissolved in dilute sulphuric 
add, and worked up exactly as already described on page 166 ; after 
careful fractionation of the acid, the portion boiling at 163 — 164° was 
converted into the silver salt and analysed. 

Digitized by VjOOQIC 


0-2044, on ignition, gave 01134 Ag. Ag«: 55*47. 

0-2154 gave 0-1900 CO^, 0-0682 H^O, and 01198 Ag. C = 2410 ; 

H = 3-52; Ag:^ 55-61. 
CsHf'OOOAg requires C = 24-61 ; H = 3*59 ; Ag»55-38 per cent. 

The remainder of the acid (b. p. 152 — 158^) was converted into the 
anilide which crystaUised from light petroleum (b. p. 80 — 100°) in 
glistening, feathery needles melting at 104 — 105° (compare Trans., 
1898, 73, 34). 

0*1296 gave 10 c.c. moist nitrogen at 20° and 762 mm. N=:8-84. 
CaHY'OO'NH'C^Hfi requires N = 8-59 per cent. 

The acid (50 per cent, yield) produced by the oxidation of isopropyl- 
acrylic acid was, therefore, isobutyric acid, and the oxidation had taken 
place in the manner expected. 

Chsmical LaboratobTi 

St. Thomab's HosprrAL. 

XX. — Some Halogen Derivatives of Acetone Dicarb- 
OQcylic Acid. Part I, 

By Fbxdebick W. Dootsok, M.A. 

The extremely reactive nature of the hydrogen atoms of acetone- 
dicarboxylic acid, C0(GH2* COOn)^^, has already been shown by a 
long series of experiments, but, so far, no attempt to replace them by 
halogens has been recorded. This is the more remarkable, since 
halogen substitution products could not fail to lend themselves to 
reactions which would yield derivatives not easily obtainable by other 
means. The present paper is intended as a preliminary note of 
investigations undertaken in this direction. 

Acetonedicarboxylic acid in aqueous solution reacts readily with 
chlorine and bromine, evolving carbon dioxide, and yielding an oil 
which attacks the eyes, and which, from its general properties, is 
probably a mixture of the halide derivatives of acetone. The inter- 
action of these halogens with the ethylic salt, however, proceeds 
smoothly and with but little decomposition, yielding ultimately ethylic 

Ethylic MrachloraeeUmedic(vrh(>xyl(Ue is easily obtained by passing a 
stream of dry chlorine into ethylic acetonedicarboxylate ; the latter, 
prepared by the action of hydrogen chloride on an alcoholic solution 
of the acid {AnndUn^ 1891, 261,177), is sufficiently good for the purpose 
without further purification. Hydrogen chloride is freely evolved, 

Digitized by VjOOQIC 


and the liquid becomes hot ; the operation is continued until chlorine 
is no longer absorbed, when the contents of the flask are found to 
have increased in weight by about 65 per cent., the reaction as it 
approaches completion being facilitated by heating on the water-bath. 
The heavy, pale yellow, oily product is fractionated under reduced 
pressure, when the greater part of the liquid distils over between 
180° and 183° (16 mm. pressure) ; this fraction is nearly pure ethylic 
tetrachloracetonedicarboxylate. A readier method of purification 
is found in fractional crystallisation. On cooling in a freezing 
mixture, a nearly solid mass is obtained which, after being drained by 
the aid of the filter pump, is strongly pressed between folds of bibulous 
paper ; the crystals thus obtained are practically pure. 

Ethylic tetrachloracetonedicarboxylate is very soluble in cither, 
benzene, alcohol, chloroform, carbon bisulphide, and light petroleum, 
but insoluble in water. On recrystallisation from alcohol, it is 
obtained in large, lustrous, colourless, rectangular plates or leaves, 
which melt constantly at 30 — 30*5° (uncorr.). 

0-3660 gave 04220 COg and 0-0947 H^O. H = 2-87 ; C = 31-44. 
0-3455 „ 0-5860 AgOl. 01 = 41-87. 

CgHjoOl^Og requires =» 31-81 ; H = 2-94 ; 01 = 41 68 per cent. 

Since all the hydrogen atoms of the acid radicle have been dis- 
placed by chlorine, the constitution of this substance is without doubt 
represented by the formula 00(00l2- 00002H5)2. 

So far this ethereal salt has resisted all attempts to saponify it. 
Odd, dilute aqueous potash seems to be without action ; aqueous soda 
or potash, on warming, liberates alcohol, but the acid at the same time 
decomposes with the formation of oxalic acid, recognised by the 
insolubility of the calcium salt in acetic acid, by a melting point 
determination (104° uncorr.), and by a titration of the crystallised acid. 
Sodium ethoxide in alcoholic solution behaves in a similar manner, 
whilst water, hydrochloric acid, and hydrobromic acid at 200^ are 
without appreciable effect. 

Action of AhohoUc Potash, — ^If a cold alcoholic solution of potash 
be slowly poured into an alcoholic solution of the ethylic salt, the 
contents of the flask being kept cold, no potassium chloride separates, 
but af fcer a short time a copious crop of colourless crystals is obtained ; 
these are exceedingly soluble in water, but only slightly so in alcohol. 
An analysis showed them to be the potassium salt of diMorofnaiUmiG 
acid (Oonrad and Bruckner, Ber,, 1891, 24, 2993). The following 
numbers were obtained on analysis. 

0-2297 gave 0-1587 K2SO4. K = 3098. 
0-2159 „ 0-2480 AgOL 01 = 28-42. 

OOl2(OOOK)2 requires K = 31-32 ; 01 = 2842 per cent. 

Digitized by VjOOQIC 


The mother liqnor from these crystals was diluted, acidified, and 
extracted with ether, and the ethereal solation washed with water, 
dried over calcium chloride, and evaporated, when a liquid with a 
strongly acid reaction was left. A consideration of the action of 
ammonia on ethylic tetrachloracetonedicarboxylate, described l)elow, 
ItoVes no room for doubt that this liquid is diMoraceUc acid. Hence 
the reaction that takes place is represented by the equation 

OqCCljCOOCjHj)^ + 3K0H = COl2(COOK)2 + 

CHClj- COOK + 202H5- OH. 

AeHon of Aqveaus Ammoiwa. — Ethylic tetrachloracetonedicarb- 
oxylate dissolves readily in aqueous ammonia with evolution of heat, 
and, on standing, a bulky crop of large, colourless crystals separates, 
these, on being recrystallised several times from hot water, are 
obtained in thin, rhombic leaves or plates, which melt constantly at 
204 — 205^ (unoorr.), are very soluble in alcohol and hot water, and 
moderately in ether, light petroleum, and benzene. From the melt- 
ing point, although this is somewhat higher than that found by 
Conrad and Bruckner (loe. cU.), and on analysis, this substance was 
identified as diehloromalonamide. 

0-2083 gave 0-1596 COg and 00472 H^O. C = 20-89 ; H = 2-51. 
0-1464 „ 20-4 c.c. nitrogen at 18° and 768 mm. N = 16-57. 
0-1470 „ 21-1 c.c. „ 19-5° and 750 mm. N« 16-58. 

0-1624 „ 0-2740 AgCl. CI -41-75. 
CCl2(CONH5)2requiresC = 2108; H=.2-35;N = 16-40; Cl = 41-42per cent. 

The mother liquor from the dichloromalonamide, on concentration, 
yielded a crop of crystals which, after recrystallisation from hot water, 
were obtained in short, thick, colourless needles with ill-defined ends ; 
these crystals, which are very soluble in alcohol and hot water, and 
moderately so in cold water, melted sharply at 98 — 99° (uncorr.). 
This, together with a nitrogen determination, was sufficient to identify 
them as dichloraeetamide (Hantzsch and Zeckendorf, Ber., 1887, 20, 

0-2000 gave 19-5 c.c. nitrogen at 17° and 762 mm. N » 11*35. 
CHCI2CONH2 requires N =» 1 1 06 per cent. 

From these results, it follows that the interaction of ethylic tetra- 
chloracetonedicarboxylate and ammonia must be represented by the 
following equation. 

CO(CCljCOOC2H5)2 + SNHg = CCl2(CONH2)2 + 

CHClg- CONH2 + 2C2H5- OH. 

From the above experiments, it appears that tetrachloracetone- 
dicarboxylic acid, if capable of existence at all, must be a very 

Digitized by VjOOQIC 


unstable compound towards alkalis, and that the displacement of 
hydrogen atoms by chlorine exercises a remarkably modifying influence 
on the nature of the products of decomposition, whilst the chlorine 
atoms themselves show an unexpected degree of stability. An 
attempt to obtain condensation products by the interaction of the 
dipotassium derivative of ethylic acetonedicarboxylate and ethylic 
tetrachloracetonedicarboxylate in alcoholic solution was unsuccessful, 
there being no separation of potassium chloride evenaf ter long-continued 

The action of bromine on ethylic acetonedicarboxylate is very 
similar to that of chlorine, hydrogen bromide being freely evolved 
and the liquid remaining almost colourless. After adding excess of 
bromine and heating for some time on the water-bath, the contents 
of the flask were shaken with dilute sodium carbonate solution until 
the colour was discharged, extracted with ether, and the ethereal 
solution washed several times with small quantities of water. After 
drying and evaporating the ether, a pale yellow oil was left^ which 
became very viscid in a freezing mixture, but could not be obtained 
crystalline. The reactions of this oil, although not identical with, 
are very similar to those of ethylic tetrachloracetonedicarboxylate, 
and will be described in a future communication. 

Uniyebsitt Chemical Labobatory, 

XXL — Action of ChlorosuVphonic Acid on the 
Paraffins and other Hydrocarbons as a means 
of Purifying the Normal Paraffins, 
By Sydney Young, D.Sc., F.R.S. 

In a paper by Dr. F. E. Francis and myself on the action of fuming 
nitric acid on the paraffins and other hydrocarbons (Trans., 1898, 78^ 
932), it was mentioned that Aschan had recently observed (^er., 1898, 
31, 1801) that di-isopropyl and isopentane are rapidly attacked by 
chlorosulphonic acid at the ordinary temperature. Aschan expressed 
the intention of continuing this investigation, and we suggested that 
it would be interesting to know whether chlorosulphonic acid, like 
fuming nitric acid, acts le^s energetically on the normal paraffins than 
on isoparaffins and other hydrocarbons that contain a :CH-group. 

It occurred to me that, if this were so, it was not improbable thali 
chlorosulphonic acid might be a useful reagent for removing small 
quantities of isoparaffins and methyl derivatives of the polymethylenes 
from the normal paraffins separated by fractional distillation from 

Digitized by VjOOQIC 


petroleam. I, therefore, wrote to Prof. Aschan to ask him whether 
he had made any observations in this direction ; he informed me» in 
reply, that he was chiefly interested in the '^ mechanism " of the 
reaction, and that he wished to continue the investigation from this 
point of view himself, but he very kindly expressed his willingness 
that I should compare the behaviour of the normal and isoparaffins 
towards chlorosulphonic acid, and that I should make use of the 
reagent as a means of purifying the paraffins. 

As a preliminary experiment, I took equal small quantities of pure 
normal hexane and of isohexane, and added to each about its own 
volume of chlorosulphonic acid. The difference in the action was ex- 
ceedingly marked ; the isohexane became warm, and in a few moments 
bubbles of gaseous hydrogen chloride were rapidly evolved ; the normal 
hexane, on the other hand, did not rise perceptibly in temperature, 
and it was only after some time that hydrogen chloride was given off. 
After standing all night, the isohexane had entirely disappeared, but 
about half the normal hexane was still left ; the chlorosulphonic acid 
had changed to a dark, tarry oil, much blacker and thicker in the case 
of isohexane. 

Again, pure normal octane from octylic iodide was far less rapidly 
attacked than normal octane separated from American petroleum by 

In order to find whether normal paraffins from petroleum could be 
purified by means of chlorosulphonic acid, specimens of normal heptane 
and of normal octane freed from aromatic hydrocarbons were treated 
with about one-fifth of their volume of the acid and left in contact 
with it for two or three days, until the reaction had ceased. The 
dark oil was in each case separated, and the remaining hydrocarbon 
was shaken repeatedly with strong sulphuric acid until, on dilution, 
the acid gave no turbidity with water. The hydrocarbon was washed 
with water, then shaken with caustic soda, and, lastly, with water 
again ; it was dried with phosphorus pentoxide and distilled twice 
from an ordinary distillation bulb. The sp. gr. was then determined, 
and the treatment with chlorosulphonic acid repeated. There was 
not a sufficient quantity of either paraffin, and they were both 
too impure, to begin with, to allow of the treatment being continued 
until the sp. gr. had become constant, but the following table will 
show that the normal octane was obtained nearly pure and that con- 
siderable progress had been made in the purification of the normal 


Digitized by VjOOQIC 


Normal Octane, 

Sp. gr. 074^ B. p. (760 mm.). 

Pure from octylic iodide 0'71848 125-75^ 

Specimen from petroleum before "t ^,,- .«, f collected between | 

treatment with chlorosulphonicacid j ^''^^^ \ 125*2° and 127'0° J 

After first treatment 0-7308 125*2 —127-3° 

„ second „ 07224 1257 —1261° 

„ third „ 0-71956 125-75—125-95° 

Normal Heptam^, 

Sp. gr. 074^ B. p. (760 mm.). 

Pure from P{nu8 sabiniana 070048 98-45° 

Specimen from petroleum before 1 r^.tjirvn j collected between | 

treatment with chlorosulphonicacid J | 98° and 102° j 

After first treatment 0*7319 99-1— 100*3° 

„ second „ 0-7223 99*1— 996° 

The normal heptane must have contained a little isoheptane and a 
large amount of methjlhezam ethylene, and the normal octane a little 
iso-octaneand a larger quantity of, presumably, dimethylhexamethylene. 
The normal octane was less impure than the normal heptane, and 
there was more of it (about 50 grams before treatment with chloro- 
sul phonic acid) ; it was, therefore, possible to carry the purification 
farther, and there can be little doubt that if the quantity had been 
larger the impurities might have been completely eliminated. 

Miss Fortey has also found that impure hexamethylene from Galician 
petroleum, boiling at a somewhat lower temperature than the pure 
substance, and, therefore, containing methylpentam ethylene and 
possibly some low boiling heptane, has both its sp. gr. and its boiling 
point raised by treatment with chlorosulphonic acid. 

When chlorosulphonic acid is added to benzene, a violent action 
takes place, and torrents of hydrogen chloride are evolved. 

The results obtained are sufficient to show that chlorosulphonic 
acid in the cold, like fuming nitric acid at a higher temperature, acts 
far less energetically on the normal paraffins than on the isoparaffins, 
the methyl derivatives of the polymethylenes or on benzene, also less 
energetically on the polymethylenes than on their methyl derivatives. 
Generally, chlorosulphonic acid acts more energetically on hydro- 
carbons that contain a :CH-group than on others. 

There is, however, some action on the normal paraffins, and con- 
sequently there is considerable loss in the process of purification, just 
as there is with fuming nitric acid. In order to obtain the pure 

Digitized by 



normal paraffins from petroleum, it would be best, before subjecting 
tbem to the action of chlorosulphonic acid, to separate them more 
completely from other hydrocarbons by fractional distillation than 
was done in the ease of normal heptane or normal octane. 

Ukivbrsity College, 


XXII. — Oxidation of Sulphocamphylic Acid. 

By W. H. Pkbkin, jun. 

When sulphocamphylic acid is oxidised at 0° with potassium perman- 
ganate, the principal product is a beautifully crystalline substance 
which melts at 254^ and, on analysis, gives numbers agreeing with 
the formula Cj8^22^7 > ^^ 100^, it loses one molecule of water, which 
is probably water of crystallisation, and is converted into a substance, 
C^gHg^O^ which has been named " dicampherylic add" 

Dicampherylic acid is a dibasic acid, as is shown by the analysis of 
its silver salt, O^gH^gAggOg, and of its methylic salt, OjgH|g(CH3)20g, 
its formula may, therefore, be written Oj0QjgO2(COOH)2. The acid, 
furthermore, contains two ketonic groups, since, with hydroxylamine, 
it yields a diozime, Ci4Hig(CINOH)2(COOH)2, and is, by reduction 
with sodium amalgam, converted into tetrahydrodicampheryUo ctctd, 
C^4Hjg(CH'OH)j(COOH)2 ; its formula may, therefore, be developed to 
Ci4Hjg(CO)2(COOH)2. The most interesting property of dicampherylic 
acid is its decomposition by concentrated sulphuric acid at 85^, when 
it is converted almost quantitatively into an acid of the formula 

^18^20^6 = 2C9H10O8. 

The determination of the nature of the acid, CgHi^Og, was found to 
be a problem of considerable difficulty, but its constitution was ultim- 
ately established by the consideration of the following facts. 

This acid is a monobasic acid, giving well characterised salts, of 
which the silver aaltf C^Hf^AgOg, the beautifully crystalline methylic 
aalt, G^H^iOR^Y)^ and the ethylio salt, Cja.Q{0^'H.^)Osy were analysed, 
and since, on treatment with acetic anhydride, it is converted into an 
acetyl campoundy C9Hg(OC2HgO)02, it is evident that it is a hydroxy. 
carboxylic acid of the composition CgHg(OH)*COOH. 

When exposed to the action of bromine vapour, the acid, by substi- 
tution, yields a dibroTno-derivative, CgHgBr2(0H)-C00H, in which the 
bromine is so firmly united that boiling with potash does not eliminate 
it, a behaviour which indicates that the acid belongs to the aromatic 

Digitized by VjOOQIC 


series. This supposition is confirmed by the fact that nitric acid con- 
verts the acid into a dinitro-derivative, CgHg(N02)2(OH)-COOH, an 
interesting substance which behaves like a dibasic acid, since its 
silver salt, which crystallises in dark red needles, has the formula 
CgH^AgjNjOy, and is evidently produced by the substitution, not only 
of the hydrogen of the carboxyl group, but also that of the hydroxyl 
group, by silver, that is, its formula may be written 

This behaviour is quite usual in aromatic hydroxy-acids containing 
two nitro-groups. Thus, for example, 3 : 5-dinitrohydrocumaric acid, 
OH- CgH2(NOg)2-CH2-CH:3-OOOH, yields a dibasic silver salt crystal- 
lising in dark red needles, and 1:3: 5-dinitroparahydroxybenzoic acid, 
OH*CgH2(N02)j'COOH, a dibasic salt crystallising in brown needles. 
From these results, it was evident that the acid O^H^qO, was either a 
hydroxydimethylbenzoic acid, OH- CgH2(CH3)2* COOH, or a hydroxy- 
ethylbenzoic acid, OH* CgH3(02H5)'COOH ; it could not be a hydroxy- 
hydrocumaric acid, OH'CgH^-CHg-OHg'COOH, since the three iso- 
meric (o-, m-, p-) forms of this acid are known, and melt respectively 
at 83°, IIP, and 129°, whereas the acid CgHj^Os melts at 204° In 
order to determine whether it contained an ethyl group or two methyl 
groups, it was decided to endeavour to oxidise the acid. Before doing 
this, however, it was necessary to protect the hydroxyl group by con- 
verting it into a methoxy-group, and this was readily accomplished by 
treating the methylic salt of the acid with sodium methylate and 
methylic iodide in the usual way. The oily methylic salt, 

thus formed yielded, on hydrolysis, the corresponding methoxy-acid, 
OCHg'CgHg* COOH, and this, on oxidation with permanganate, was 
almost quantitatively converted into an acid of the formula 
OCH3-C7Hg(OOOH)2, showing that the original acid, CqEL^qO^, must 
have contained two methyl groups, one of which had been converted 
into COOH during the oxidation ; in short, the acid CgHj^Og is a 
dimethylhydroxybenzoic acid, OH-CgH2(CH3)2'COOH. 

On heating the methoxydibasic acid, OCH3'C7H4(OOOH)j, with 
hydriodic acid, the corresponding hydroxy-acid, OH*C7Hg(COOH)2, 
was obtained as a very sparingly soluble, crystalline powder melting 
at 283°, and giving, in alcoholic solution, an intense reddish-violet 
coloration with ferric chloride. As these properties are the same as 
those of the hydroxymethylterephthalic acid of the formula 




Digitized byCjOOQlC 

perkin: oxidation of sulphocamphylio acid. 177 

described by Jacobsen and H. Meyer {Ber., 1883, 16, 191), the 
author at first thought that the acids were identical (compare Froc., 
1893, 9, 110). 

The experiments up to this stage were completed in 1893, and 
since then, during the course of a long investigation into the pro- 
perties of sulphocamphylic acid, some new facts were discovered which 
made it improbable that this formula could be correct. 

The subject was therefore further investigated, and after much labour 
the relative positions in the hydrozydimethylbenzoic acid, CgH^^Og, and 
also in the hydroxymethyldicarboxylic acid, OH-CgHj(CH3)(COOH)2, 
produced by its oxidation have, as the author believes, been clearly 

In the first place, the acid, CqBL^qOq, which melts at 204^ and gives 
no colour with ferric chloride, is isomeric with the already known 
hydroxydimethylbenzoic acids, as will be seen from the following 
table, in which the numbers refer to the groups taken in the order 


' ^ M. p. 

oriho'HydriKBypM'aicylie acid 1:2:5:4 
i^riho-HydrooBymeaitylenic acid 1 ! 3 : 4 : 5 
parorMydraxymesitf/lenic acid 1:3:2:5 
a-Hydroocyxylic add 1:3:1:4 

P' u » 1:4:2:3 

T » » 1:4:5:2 

The careful investigation of the acid CgHj^Os, melting at 204^, has 
clearly established the following points. 

1 . J%e Hydroxyl Group i$ in the Afetorpoai^ion relatively to the Carhoxyl 
Grotip. — That the acid is not an orthohydroxy-acid is clearly indicated 
by the fact that it gives no colour with ferric chloride ; that it is 
neither an or^hydroxy- nor a ^xirahydroxy-acid is shown by the fact 
that it may be heated with concentrated hydrochloric acid or hydriodic 
acid at 200 — 220^ without decomposition, whereas ortho- and para- 
hydroxy-acids are readily decomposed into the corresponding phenols 
and carbonic anhydride by this treatment. 

^ 2. The Acid w derived from the Xylmud, [(CH3)2 : OH = 1 : 2 : 6].— 
When the calcium salt of the acid CgH^^Og is distilled with lime, it 
shows the behaviour characteristic of the calcium salts of meta- 
hydroxy-acids, namely, decomposition takes place with great difficulty, 
and only at an exceedingly high temperature. The distillate contains 
a xylene], which crystallises from water in long, colourless needles, 
and the aqueous solution of which gives, with ferric chloride, a dis- 
tinct, although not intense, bluish-violet coloration ; it melts at 75^ 
and distils constantly at 218^ The identity of this xylenol was, with 
difficulty, established in the following way. All the six possible 

Digitized by VjOOQIC 


FeCl,, bluish-violet. 


FeCl,, intense blue. 


FeClg, brown. 





perkin: oxidation of sulphocamphylic acid. 

xy lends, CgH3(CH3)2*OH, have been prepared and carefully in- 
vestigated, 80 that their properties are well known; and in order 
to clearly show how the constitution of the zylenol from the acid 
CgHjoOg was proved, a short table of the properties of the 
zylenols is appended, the numbers referring to the groups in the 
order CHg : CH3 : OH. 



b. p. 219 -S* 

m. p. 68* 

Fed,, no colour. 



„ 218 

„ 76 

,, violet 



„ 211-6 

,, 76 

„ no colour. 



» 226 

» 66 

„ not stated. 



„ 211-6 

» 26 

„ blue. 



„ 212 

„ 49 

„ not stated. 

Three of these zylenols, namely, Nos. Ill, lY, and Y, were obtained from 
Schuchardt, and after very careful purification found to be different 
from the author's zylenol, as, indeed, the melting and boiling points 
would indicate. Since, again, both the melting point and boiling point 
of No. YI differ markedly from those of the author's zylenol, there re- 
main only Nos. I and II as possibly representing its constitution, and of 
these, No. II is the more probable, since this melts and boils at ezactly 
the same temperatures, whereas No. I has a lower melting point, and is 
stated to give no coloration with ferric chloride. As, however, very 
slight traces of impurity considerably depress the melting point of the 
zylenols, and as the colour with ferric chloride is not pronounced in the 
case of any one of them, it was necessary to obtain further evidence 
before deciding this question of identity. Fortunately, there is a wide 
difference in the melting points of the tribromo-zylenols derived from 
Nos. I and II, that from the former melting at 166°, whereas that from 
the latter melts at 184°, and on trying the ezperiment with the author's 
zylenol it was found that its tribromo-derivative melted sharply at 
184° ; there can, therefore, be no doubt that the constitution of this 
zylenol is represented by the formula Cq'K^{CH^)2'0^'^ 1:2:6. 

Since, then, the acid, On*C0H2(CH3),* COOK, from which this 
zylenol was obtained is a metahydrozy-aoid, it can only have the 
formula OeH2(CH3)2(OOOH)-OH= 1 : 2 : 4 : 6 ; that is, it is metahydr- 
axypouraaaylic (leid^ and closely allied to arthohydroxi/paraasylic ctcid, 
[1:2:4:6], which Renter (Ber,^ 1878, 11, 30) prepaxed by fusing 
pseudocumenesulphonic acid with caustic potash (compare Jacobsen, 
Ber., 1899, 12, 436). This is the second occasion on which a 
derivative of pseudocumene has been obtained indirectly from sulpho- 
camphylic acid, as Koenigs and Meyer {Ber., 1894, 27, 3468, [com- 
pare Trans., 1898, p. 840) found previously that isolauronic acid was 
readily converted into parazylic acid by the action of sulphuric acid. 

The hydrozy-dibasic acid, produced indirectly by the ozidation of 

Digitized by VjOOQIC 


metahydrozyparaxylio acid as described above, is evidently an ortho- 
hydrozy-acid, since its alcoholic solution gives an intense blue 
coloration with ferric chloride; it can, therefore, only be a 
hydrozymethylterephthalic acid of the formula 

0H-0eHj(CH3)-CX)0Ha [CHj-COOH : OH : COOH = 1:3:5:6], 

and the difference between this formula and that originally assigned 
to this add (see p. 176) is simply in the position of the hydrozyl 

Unfortunately, like so many other instances in which benzene com- 
pounds have been obtained from cafnphor and its derivatives, the 
formation of metahydrozyparazylic acid throws very little light on the 
constitution of dicampherylio acid, from which it is so readily pro- 
duced, and although it would be easy to suggest a formula for 
the latter acid, the author prefers not to do so until further experi- 
mental data are forthcoming. 

With regard to the formation of such benzene derivatives as xylene, 
cymene, pseudocumene, carvacrol, acetylortho-xylene, and others from 
camphor, and of derivatives of paratoluic acid and parazylic acid from 
camphoric acid, the author entirely shares the opinion, somewhat 
differently expressed by Ascban,* namely, that until we understand 
the extraordinary mechanism of the production of these benzene 
derivatives, their formation is of little value in determining the 
constitution of the substances from which they are derived. 

The author is much indebted to Messrs. B. Prentice, J. L. 
Heinke, F. H. Lees, and others for their valuable assistance in the 
experimental part of this investigation. 

The author also wishes to state that the heavy cost of the very large 
amount of material used in this research was to a very considerable 
extent covered by repeated large grants from the Government Grant 
Fund of the Koyal Society. 


OxidtUian of Sulpkocamphylic Acid. Formation of Dic(vmpherylic Add, 

CisHjoOe + HjO. 

In carrying out this oxidation, sulphocamphylic acid in quantities 
of 60 grams was dissolved in about 1 litre of water, the solution 
neutralised with potassium carbonate, and transferred to a flat 
porcelain basin in which a turbine was fitted. The solution was 
cooled to (P by adding powdered ice, and a cold saturated solution 

* Straetur nad steraochenuache Sta<li«ii in der Camphergrnppe ; Helsingfors, 
1895, p. 17. 

Digitized by VjOOQIC 


of potassium permaDganate run in slowly until the colour, which 
instantly disappears at first, just remained permanent, care being 
taken that the whole was well stirred and the temperature kept below 
2° during the operation. The product from several such oxidations 
was filtered, the filtrate and washings of the manganese precipitate 
evaporated to a small bulk, and the concentrated yellow solution 
acidified and allowed to stand overnight; the yellow, crystalline 
precipitate which had separated was collected, washed with water, 
and purified by repeated crystallisation from dilute acetic acid ; the 
magnificent, lemon-yellow prisms thus obtained, after standing for 
some days on a porous tile in contact with air, gave the following 
results on analysis. 

01216 gave 0-2767 00^ and 00712 HgO. = 61-83; H = 6-60. 

01410 „ 0-3190 COj „ 00806 HjO. = 6170; H = 6-36. 

01524 „ 0-3448 OOj „ 0-0894 HjO. = 61-70; H = 6-61. 

OigHjjOy requires = 61 71 ; H = 6-28 per cent. 

These crystals, when heated for a long time at 100^ gradually lose 
water, a change which takes place rapidly at 120 — 130°. 1 '0952 grams 
of the substance, dried in the air, lost 0-0610 gram at 100", or 5 '47 
per cent, whereas the calculated loss, supposing O^gEEgs^r ^ ^ 
converted into O^gH^oO^ at 100°, is 5-14 per cent. An analysis of the 
substance, dried at 120° until of constant weight, gave the following 

0-1496 gave 0-3571 00^ and 00851 Kfi. = 6510; H = 6-32. 
OigHgoOg requires = 6506 ; H = 6-02 per cent. 

The molecular weight of this substance was determined by the 
cryoscopic method, when it was found that 0*8010 gram, dissolved in 
36-1 grams of acetic acid, depressed the melting point 0*26°, this 
corresponding with a molecular weight 332, whereas the molecular 
weight of OigHj^jOg is 332 and of OigHj^Oy 350. 

On long standing in an uncorked flask, the solution used in this 
determination deposited a mass of slender, colourless needles which 
looked like threads of asbestos and quite unlike the hard prisms in 
which the substance OigH^^O^ usually crystallises. An analysis 
showed, however, that the crystals had the composition O^gHjgO^. 

0-1028 gave 0-2328 00^ and 0-06 H^O. = 61*76 ; H = 6-48. 
OigHjaOy requires = 61*71 ; H = 6*28 percent. 

These results, taken together with those described in the following 
pages, show that the product of the oxidation of sulphocamphylic acid 
is a substance of the formula O^gH^oO^, which crystallises with water 
in prisms of the composition O^gHg^O^. 

Digitized by VjOOQIC 


This compound, which has been named dicampherylic acid, melts at 
about 254^, and is readily soluble in alcohol, acetone, acetic acid, and 
ethylic acetate, but only very sparingly in benzene; it is almost 
insoluble in cold water, although, in a finely divided state, it dissolves 
appreciably in boiling water, and separates on cooling in almost 
colourless, prismatic crystals. It dissolves readily in dilute sodium 
carbonate, and, although the well-cooled solution does not decolorise 
permanganate, it does so readily on warming. 

Dicampherylic acid dissolves readily in warm, concentrated nitric 
acid, and, on boiling, oxidation takes place so slowly that even after 
10 minutes the addition of water causes a large quantity of the 
substance to separate unchanged. Fuming hydriodic acid dissolves 
the finely-powdered crystals easily, and if, after boiling for a short 
time, the bulk of the hydriodic acid is distilled off, the residue, on 
mixing with water, deposits a heavy, dark brown oil which becomes 
colourless on the addition of sulphurous acid and smells like isoamylic 

Salts of Dicampherylic Acid. 

That this acid is a dibasic acid was first shown by titrating a 
weighed quantity of the anhydrous crystals with standard alkali, 
when it was found that 0*4416 required for neutralisation 0'1495 
gram KOH, or 33*8 per cent., whereas a dibasic acid of the formula 
C^gHg^Og requires 33*7 per cent. 

The silver salt was obtained as a white, gelatinous precipitate on 
adding silver nitrate to a warm, neutral solution of the ammonium 
salt, the precipitate on warming with much water for about an hour 
becomes crystalline, and is then readily filtered and washed. The 
analyses seem to show that this salt, after drying over sulphuric acid 
in a desiccator, has the formula CigHg^AgjOy, and that at 120—130° 
it loses water, and then has the formula Cj^H^gAggO^. 

The silver salt, dried over sulphuric acid in a desiccator, gave the 
following results on analysis. 

0-2064 gave 0*2833 CO2, 0*0696 H^O, and 0*0790 Ag. C = 37*66 ; 

H = 3*77j Ag- 38*46. 
0*2469 gave 0*3420 00^, 0*0746 ttjO, and 00963 Ag. « 3777 ; 

H = 3*36; Ag« 38*68. 
0*3390 gave, on ignition, 0-1 302 Ag. Ag » 38*41. 
CigHj^gjO^ requires C » 38*29. H = 3*66 ; Ag = 38*29 per cent. 

This salt may be heated at 120 — 130° without decomposition, and 
after its weight had become constant it was analysed. 

Digitized by VjOOQIC 


0-2260 gave 03223 OO^, 0-0681 H^O, and 0-0887 Ag. = 38-89 ; 

H=3-35; Ag = 39-24. 
0-22, on ignition, gave 0-0860 Ag. Ag = 3909. 
CigHigAgjOj requires = 3956 ; H = 3-30; Ag= 39*56 per cent. 

A neutral solution of the ammonium salt of dicampherylic acid 
gives a heavy, white precipitate with lead acetate, but no precipitate 
with barium, calcium, or zinc chlorides ; copper acetate gives no pre- 
cipitate in the cold, but on boiling a light blue, crystalline, copper salt 

Meihylic Diccbmpherylate, CiqH.^q{C'H^}20^. — ^This salt was prepared 
by two different methods. 

1. The dry silver salt of the acid was digested for 2 hours with ether 
and excess of methylic iodide, and the ethereal solution filtered ; the 
residue was extracted repeatedly with ether, and the combined 
ethereal solutions evaporated to a small bulk. On standing, beautiful 
crystals separated which, after crystallisation from benzene, melted at 

2. The acid, dissolved in methylic alcohol, was mixed with con- 
centrated sulphuric acid and heated to boiling on a water-bath for 
a few minutes ; on cooling, glistening crystals separated, which, after 
collecting, and recrystallising from benzene, melted at 226 — 227°. 
This substance, which, like all the derivatives of dicampherylic acid, is 
very difficult to bum, was analysed, with the following results. 

I. 01692 gave 0-4120 OO2 and 01040 HjO. = 66-42; H = 6-83. 
n. 01652 „ 0-4011 OO2 „ 0-0996 HjO. = 66-22; H = 6-69. 
III. 0-1663 „ 0-4027 OO2 „ 01007 HjO. = 66 04; H = 6-73. 
Oi8Hi8(OH8)20^ requires = 66 66 ; H = 6-66 per cent. 

In No. I, the methylic salt used was prepared from the silver salt, 
and in Nos. II and III from the acid by treatment with methylic alcohol 
and sulphuric acid. 

Action qf ffydroacf/lcmUne on Dicampherylie Acid, 

In studying the action of hydroxy lamine on dicampherylic acid, the 
pure acid (2 grams), dissolved in dilute potash (6 grams), was mixed with 
hydroxy lamine hydrochloride (2 grams), and acidified after 12 hours, 
when the nearly pure oxime separated in white flocks.* The whole 
was extracted with much ether, the ethereal solution rapidlyf dried 

* In this form, the oxime is readily soluble in ether, and also in warm water, bat 
after crystallisation from ether, it becomes almost insoluble, both in this solvent 
and in water. 

t If this operation is not very quickly performed, the oxime is apt to separate on 
the lumps of calcium chloride. 

Digitized by VjOOQIC 


with calcium chloride, filtered, and allowed to stand either as it was, 
or, better, after concentrating somewhat, when the ozime separated in 
crystalline crusts. These crusts, after washing with ether, gave the 
following results on analysis. 

I. 01420 gave 0-3090 CO2 and 00830 H^O. C = 5935 ; H = 650. 

IL 0-1548 „ 0-3362 00, „ 0-0914 HgO. = 59-23 ; H = 6-57. 
ni. 01699 „ 0-3663 OOj „ 0-0986 HjO. = 58-80; H = 6-44. 
IV. 01608 „ 0-3487 OOj „ 00937 H^O. = 5914; H = 6-47. 

V. 0-2602 „ 17-4 c.c. nitrogen at 18° and 758 mm. N = 7-70. 
Other nitrogen determinations gave N = 6'90, 7*30, and 7*05. 

OigHjjNaOg requires = 59-66 ; H = 607 ; N = 773 per cent. 

This substance is, therefore, the dioxime of dicampherylic acid produced 
according to the equation 

CisHjoOo + 2KH,-0H - O.gHjoO.CN-OH), + 2H,0. 
When heated in a capillary tube, the ozime turns brown at 210 — 220°, 
and gradually becomes darker as the temperature rises, but does not melt 
completely at 250°. It is almost insoluble in ether and most organic 
solvents, but dissolves in concentrated hydrochloric acid, and is re- 
precipitated unchanged when the solution is diluted with water and 
warmed; it dissolves also in caustic potash solution, and is repre- 
dpitated unchanged by hydrochloric acid. 

Action qf Acetic Anhydride on the Dioxime.-— -The pure dioxime 
dissolves slowly, but completely in cold acetic anhydride, and if the 
solution is allowed to remain in contact with air, a crystalline cake 
gradually forms ; this, after recrystallisation from dilute acetic acid, 
gave the following results on analysis. 

0-1524 gave 0-3130 OOj and 0800 HgO. = 560 ; H = 6-83. 
0-1277 „ 0-2640 OO2 „ 0-0702 H^O. = 5638; H = 6-ll. 
0-2106 „ 12-4 C.C. nitrogen at 18° and 742 mm. N = 663. 
Oj^jHjjjNjOj requires = 56-87 ; H = 6-13 ; N = 6-63 per cent. 

This substance, which is doubtless the (iceicUe of iha dioxime, 
OjgH^gNjO^O^H^O,, melts at 184°, and crystallises from dilute acetic 
acid in microscopic needles; when heated at 190 — 200°, it froths up 
and emits an odour of acetic anhydride. Dilute potash hydrolyses this 
acetate, and on acidifying and extracting with ether in the usual way, 
the characteristic crusts of the regenerated dioxime are obtained. 

0-3153 gave 22*5 c.c. nitrogen at 22° and 758 mm. N = 8-06. 
OigHjjN^O^ requires N » 7*73 per cent. 

Digitized by VjOOQIC 

184 perkin: oxidation op sulphocamphylic acid. 

Action of Phenylhydraaine on DicamfJierylic Add, 

When dicampherjlic acid (3*6 grams) is mixed with freshly-distilled 
phenylhydrazine (5 grams) and glacial acetic acid (10 grams), and the 
mixture heated on the water-bath, reaction sets in in a short time with 
evolution of gas, the liquid soon becomes dark-red, and after 20 
minutes crystals begin to separate. As soon as the separation appears 
to be complete, the whole is filtered hot on a platinum cone,* and the 
crystals washed with acetic acid and recrystallised from this 
solvent. The ruby-red crystals thus obtained are very difficult to 
bum, and for this reason several analyses were made^ with the 
following results. 

01487 gave 0'3736 COg and 00820 HgO. = 68-50 ; H = 6-13. 
0-1735 „ 0-4333 CO2 „ 00957 HgO. = 68-16; H = 6-I3. 
0-1562 „ 0-3873 OO3 „ 0908 HjO. = 6762; H = 6-46. 
01963 „ 18-5 C.C. nitrogen at 21° and 753 mm. N = 10*64. 

These analytical numbers agree best with the formula OjoHg^N^Og, 
which would be obtained by the combination of 1 mol. of dicam- 
pherylic acid and 2 mols. of phenylhydrazine with loss of water, 
Oi8H2oOe + 20eH5-NH-NH2 = C3oH8^NP5 + H20, or it is possible that 
it is the dihydrazone of O^gHgoOg crystrallising with IHjO. 

This substance dissolves in hot acetic acid or alcohol, forming yellow 
solutions which, on cooling, deposit brilliant, ruby-red crystals. When 
heated in a capillary tube, it darkens at 230°, and decomposes with 
evolution of gas and charring at 237°. It dissolves in alkalis and 
alkali carbonates, forming yellow solutions, which, on acidifying, give 
a bright yellow precipitate, consisting, apparently, of the unchanged 

Reduction qf Dicampherylic Acid. FormcUion of Tetf^hydrodieam- 
pherylic Acid, O^^'H^^fio' 

In order to investigate this reduction, pure dicampherylic acid was 
dissolved in dilute soda and treated in a flat basin with a large excess 
of sodium amalgam (5 per cent.), and after 24 hours, the whole 
was neutralised, filtered, and the filtrate acidified and extracted 
many times with ether. The ethereal solution, after drying and con- 
centrating, deposited the greater part of the product of reduction in 
crystalline flakes. These were collected, washed with ether, and 

* The mother liquors when heated with more phenylhydrazine give another 
crop of crystals. 

Digitized by VjOOQIC 


01683 gave 0-3966 COj and O'l 126 H^O. C = 6409 ; H = 7-43. 

0-1600 „ 0-3497 CO2 „ 0-1003 HgO. C = 63-58 ; H= 7-42. 

0-1616 „ 0-3816 CO3 „ 0^080 H2O. = 6444 ; H = 7-43. 

CigHg^Og requires C = 64-28 ; H = 7*14 per cent. 

The determination of the molecular weight of this acid by the boiling 
point method, using alcohol as the solvent, gave 298, 307, and 296, 
whereas the calculated molecular weight for C^^K^fiQ is 336. 

Tetrahydrodicampherylic acid melts at about 297 — 298° with 
evolution of gas. It is sparingly soluble in methylic and ethylic 
alcohols and acetic acid in the cold, and in boiling water, and is 
not readily soluble even in boiling acetic acid ; but from its solution 
in dilute acetic acid it separates, on long standing, in four-sided plates. 

Silver TetrcLhydrodicwnpherylate, — ^When a hot, dilute solution of 
the ammonium salt of the acid is mixed with silver nitrate, a 
white precipitate is formed, and if this be rapidly removed with 
the aid of the pump, the filtrate, on standing, deposits beautiful groups 
of colourless crystals of a silver salt; this, on analysis, gave the 
following numbers. 

0-2169 gave 0-3070 COj, 00826 HgO, and 00846 Ag. C = 3860 ; 
H=4-18; Ag = 39-0. 
CjgHggAggOe requires » 39*27 ; H = 4-0 ; Ag = 3927 per cent. 

Fusion of DicampheryHc Acid with. Potash. 

Fifteen grams of the pure acid was mixed with 60 grams of potash 
and a little water, and heated at about 120°, when the crystals rapidly 
dissolved, forming a dark solution, which readily crystallised owing to 
the separation of a potassium salt. On raising the temperature, the 
liquid became dark brown, and then red, and after much frothing sud- 
denly solidified at about 165°, a peculiar aromatic smell being noticeable ; 
more potash (20 grams) was then added, and the temperature gradually 
raised, until, at 250°, the frothing had entirely ceased. The melt was 
dissolved in water, acidified, and extracted 10 times with ether, the 
ethereal solution evaporated, and the viscid residue, which smelt 
strongly of fatty acids, distilled in a current of steam. The 
strongly acid distillate, when neutralised with pure calcium carbonate 
and concentrated, deposited beautiful, feathery crystals of calcium 
isobutyrate ; these were collected and analysed. 

DetermiTuUion of Water of Cryatallisation,* — 0*2426 gram, on heat- 
ing at 100° until constant, lost 00674 gram H20« 27*79. 

{OJELjO^^CsL + Q'ELfi requires H20 = 29*60 per cent. 
* Compare Croesley and Perkin, Trans., 1898, 78, 15. 

Digitized by VjOOQIC 


01752 gave 0-1107 OaSO^. Ca= 18-58. 

(08H/COO)2Ca requires Ca = 18-67 per cent. 

The remainder of the calcium salt was dissolved in water, acidified 
with hydrochloric acid, and the oily acid which separated extracted 
with ether ; the dried ethereal solution, after the ether had been re- 
moved on the water-bath, left an oily residue of isobutyric acid which 
distilled at 154—165°. 

01684 gave 0-3347 00^ and 0-1368 HgO. = 54-21 ; H = 908. 
CgH^COOH requires = 54*54 ; H = 909 per cent. 

This acid, which is undoubtedly isobutyric acid, is formed in con- 
siderable quantities during this fusion, as the amount obtained from 
15 grams of dicampherylic acid could hardly have been less than 
3 grams. 

The residue in the distilling flask, after the isobutyric acid had been 
removed by steam, contained a flocculent precipitate ; this was removed 
by filtration, and the filtrate extracted five times with ether. The 
ethereal solution, on evaporation, deposited 9 grams of an almost 
colourless, thick, oily acid, which showed no signs of crystallising, 
even after standing for some weeks over sulphuric acid. On analysis, it 
gave the following numbers. 

01535 gave 0-3364 OOj and 0-0842 H^O. = 5977 ; H = 6-ll. 
CgHioO^ requires = 59*34 ; H = 5-50 per cent. 

In the hope of obtaining this acid in a crystalline condition, it 
was converted into its lead salt, by adding lead acetate to its 
aqueous solution, and the white precipitate, which looked almost 
crystalline, was suspended in water and decomposed by hydrogen 
sulphide; the filtrate, after concentration, was extracted with pure 
ether, <Sec., in the usual way, and the residue, on being allowed to remain 
over sulphuric acid in a vacuum desiccator, left a colourless, syrupy 
acid which, on analysis, gave the following results. 

0-1753 gave 03785 OOj and 0-0953 HgO. = 5888 ; H = 6-05. 
0-1418 „ 0-3086 OOj „ 0774 H^O. = 59-35 ; H = 6-05. 
OjjHioO^ requires = 59-34 ; H = 5-50 per cent. 

The silver salt, which was obtained as a white, amorphous precipi- 
tate on adding silver nitrate to a neutral solution of the ammonium 
salt, gave the following results on analysis. 

0-2157 gave 0-2184 00^, 0-0469 HgO, and 0-1173 Ag. = 27-61 ; 

H = 2-41; Ag = 54-38. 
Oj^HgAgjO^ requires = 27*28 ; H = 2-02 ; Ag = 54*54 per cent. 

This acid, which appears to have the formula Q*jB^{GO0Ti)^, is very 

Digitized by VjOOQ IC 


readily soluble in water; when heated, it gives a yellow distillate 
which is insoluble in cold water, but dissolves on boiling, and there- 
fore very probably consists of the anhydride. The solution of the 
add in sodiom carbonate does not decolorise permanganate at ordinary 
temperatures except on long standing. 

It should be mentioned that, in one experiment on the fusion of 
dicampherylic acid with potash, a small quantity of an acid was ob- 
tained which was almost insoluble in water, and only very sparingly 
■olnble in boiling glacial acetic acid. This acid, which crystallised 
from acetic acid in needles and melted above 300^, was at first thought 
to be terephthalic acid, but this is not the case, since its methylic salt 
is a syrup. 

Metahydroocyparaxylte Acid, CgHj(CHg),(COOH)-OH [=1:2:4:6]. 

This acid, as stated in the introduction, is formed by the action of 
concentrated sulphuric acid at 85^ on dicampherylic acid. The finely- 
divided acid dissolves readily in warm, concentrated sulphuric acid, 
forming a yellow solution which, if heated at 80 — 85^, gradually 
becomes darker, the decomposition of the dicampherylic acid being 
complete in about 10 minutes. On pouring the product into water, an 
ochreous precipitate separates which is collected on the pump, washed 
well with water, dissolved in dilute sodium carbonate, boiled with 
animal charcoal, filtered, and the acid reprecipitated. The now 
almost colourless acid is further purified by recrystallisation, first 
from glacial acetic acid, and then from water. 

01462 gave 0-3483 CX)^ and 00803 H^O. C = 6497 ; H = 610. 

01402 „ 0-3360 COj, „ 0773 HjO. = 65-36 ; H = 613. 

01268 „ 0-3023 OO2 „ 0721 H^O. = 65-02; H = 6-31. 

CgHjoOg requires - 65*06 ; H = 6-03 per cent. 

Metahydroxyparaxylic acid melts at 203 — 204% and when strongly 
heated, as, for example, during the combustion, a portion chars, but the 
main quantity sublimes in beautiful, colourless crystals. It is readily 
soluble in alcohol and ether, moderately so in hot water and in glacial 
acetic acid, but only sparingly in chloroform, light petroleum, benzene, 
carbon bisulphide, and cold water ; its aqueous solution gives no colora- 
tion with ferric chloride. From glacial acetic acid, it crystallises beauti- 
fully in nearly colourless, glistening plates, which, on exposure to the 
air, become opaque and chalky, a change which takes place rapidly at 
100^ ; from water, the acid also crystallises well. Hydroxyparaxylic 
acid does not react with hydroxylamine, and is not reduced by 
sodium amalgam at the ordinary temperature, or when its solution 
in hydriodic add is boiled for some time ; at 200 — 220°, however, 

Digitized by VjOOQIC 

188 perkin: oxidation of sulphocamphylic acid. 

fuming hydriodic acid and phosphorus slowly attack it, with forma- 
tion of a small quantity of a neutral substance of phenolic nature, 
but even at this temperature most of the acid remains undecomposed. 
The great stability of the acid is furthermore clearly shown from the 
following experiment. 

1'5 grams of the pure acid was heated with 10 c.c. of concentrated 
hydrochloric acid at 230 — 235° for 8 hours ; the tube contained a 
good deal of charcoal, and the greenish liquid, which had a phenolic 
odour, was filled with long, silky needles like asbestos threads ; these 
crystals were collected and purified by dissolving in potash, filtering, 
and reprecipitating with hydrochloric acid ; the substance then melted 
at 202°, and consisted of unchanged hydroxyparaxylic acid, as the 
following analysis shows. 

01 395 gave 03320 COg and 0-0778 HgO. = 64-92 ; H = 6-19. 
CgHioOg requires C = 6606 ; H = 603 per cent. 

The amount of acid recovered in this experiment was 1 '1 grams. 

Salts of Hydroxypckraxylic Acid, 

The silver salt, CgHgAgOg, was obtained, on adding silver nitrate 
to a neutral solution of the ammonium salt, as a white, amorphous 
precipitate. It decomposes readily on heating, yielding a beautifully 
crystalline sublimate. The following results were obtained on 

01736 gave 0*2486 COg, 00548 H2O, and 0*0680 Ag. C = 3906 ; 

H = 3-61; Ag = 3918. 
0-1562 gave 0-0617 Ag. Ag = 39-50. 

CgH^AgOg requires = 3956; H = 330 ; Ag = 39-66. 

The neutral solution of the ammonium salt of hydroxyparaxylic 
acid gives no precipitate with calcium* or barium chlorides. 

Methylic Hydroxyparaayylate, OH- CgH2(CH3)2- COOCH3.— In order 
to prepare this salt, the acid (22 grams), dissolved in methylic alcohol 
(160 c.c), was mixed with concentrated sulphuric acid (40 c.c.) and 
heated to boiling for half an hour; more sulphuric acid (10 c.c.) was 
then added, the boiling continued for half an hour, and the whole 
allowed to stand overnight. The beautiful, striated, leaf -like crystals 
which separated wete collected with the aid of the pump, and, after 
washing with water and drying at 100°, were found to weigh 13*8 
grams. The brown mother liquors, on diluting with water, deposited 
a brown oil which, on shaking and stirring, slowly solidified ; this 
crude product, after washing and drying on a porous plate at 100°, 

* For a description of the calcium salt of hydroxyparaxylic acid, see p. 192. 

Digitized by VjOOQIC 


weighed 7*5 grams. Each portion was separately crystallised from 
henzene, and in this way 17 grams of the pare methylic salt was 
obtained, showing that the acid is easy to etherify. 

0-1763 gave 0-4300 COg and 01087 ttjO. C = 66*52 ; H == 6-86. 
01544 „ 0-3772 COg „ 00939 HjO. C = 66-67 ; H- 6-76. 
CioHigOj requires C= 66*66 ; H = 6-66 per cent. 

Methylic hydroxypcwaosylcUe melts at 148 — 149^ and in its other 
properties closely resembles the corresponding ethylic salt. 

Bthylie hydroxyparaxylate, OH- CoHj(CH3)2* COOCjHg.— This was 
prepared by treating the acid with ethylic alcohol and sulphuric acid 
in the same way as the methylic salt. 

The colourless needles obtained by crystallisation from light pet- 
roleum melted at 134 — 135^, and gave the following results on 

0-1309 gave 0-3270 CO, and 00891 H^O. C = 68-13 3 H = 756. 

0-1590 „ 0-3962 CO3 „ 0-1010 H^O. 0-67-94; H = 7*06. 

C^jHiPg requires = 68-04 ; H = 7*22 per cent. 

Ethylic hydroxyparaxylate is readily soluble in alcohol and benzene, 
but rather sparingly in light petroleum ; its solution in alcohol gives 
no coloration with ferric chloride. 

It is insoluble in sodium carbonate, but dissolves readily in caustic 
potash solution, and is reprecipitated unchanged on the addition of 
acid. When heated in small quantities, it distils unchanged. 

Aeetaxyparastylic Acid, G^lS.j^CR^)j^OQ^fiyOOQiR. 

In order to prepare this substance, hydrozyparaxylic acid (1 gram) 
was boiled with acetic anhydride (10 grams) in a reflux apparatus for 
half an hour, and the product, when cold, was shaken with water 
until the excess of anhydride had been removed. 

The acetyl compound, which separated as a crystalline cake, was 
purified by recrystallisation, first from methylic alcohol and then from 

0-1556 gave 0*3612 COg and 00810 H^O. = 63*30 ; H == 5-78. 

0-1642 „ 0-3805 COj „ 0-0870 HjO. C==6319 ; H = 5-88. 

C11H12O4 requires 0=* 63-46 ; H=*5-77 per cent. 

Acetoxyparaxylic acid crystallises in colourless, stellate groups, melts 
at 141 — 142^, and when rapidly heated in small quantities it distils 
with only very slight decomposition. It is readily soluble in benzene, 
alcohol, and ethylic acetate, but almost insoluble in light petroleum ; 
it dissolves with difficulty in, and is only very slowly hydrolysed by, 

VOL. LXXV. Digitized by Q30gle 


boiling water, but is readily decomposed by boiling with potass- 
ium carbonate solution, and, on acidifying, pure hydroxyparazylic 
acid separates. 

Dinitrohyd/raocyparaxylic Add, ^^ I \k[(\. 


When pure hydroxyparazylic acid is added^ in small quantities at 
a time, to strong nitric acid, it hardly dissolves, and no change takes 
place until the mixture is gently warmed, when suddenly a vigorous 
action sets in, the substance passes into solution, and each fresh quantity 
added dissolves with a slight hissing sound and abundant evolution of 
red fumes. In a short time, beautiful, yellow crystals separate, and 
after cooling and mixing with an equal bulk of water, further pre- 
cipitation takes place. The yellow crystals were collected, washed 
with water, recrystallised from this solvent and analysed, with the 
following results. 

01818 gave 0-2838 COg and 0'0562 H2O. C = 42-57 ; H = 3-43. 
01637 „ 0-2535 002 „ 0-0496 H^O. = 42-23 ; H = 336. 
0-1996 „ 19-2 c.c. nitrogen at 22*^ and 764 mm. N= 10-81. 
0H-0e{0H3)2(N02)2-000H requires = 42-19; H = 3-12; N= 10*94 per cent. 

Dinitrohydroxyx)araxylic acid is readily soluble in alcohol, acetone, 
ethylic acetate, and acetic acid, moderately so in hot water, but only 
sparingly in cold water, benzene, and chloroform. It crystallises from 
water in glistening, yellow plates, and dissolves in dilute sodium 
carbonate, forming a deep orange solution. When heated in a capillary 
tube, it darkens at 195°, and decomposes rapidly at 203 — 205° with 
evolution of gas. 

SaiUs of Dinitrohydroxypwaxylic Add. 

Tlie Silver Salt, 0Ag*0^(0H3)2(N02)2*000Ag.— When silver nitrate 
is added to a moderately concentrated and slightly alkaline solution of 
the ammonium salt of the acid, no precipitate is produced at first, but, 
on standing, a splendid, dark purple, crystalline silver salt separates, 
which, under the microscope, is seen to consist of fern-like groups of 
needles. As this salt appeared to be explosive, the silver was deter- 
mined by heating the substance with nitric acid and hydrochloric acid 
in a sealed tube at 180° and weighing the silver chloride formed. 

01466 gave 0-0892 AgOl. Ag = 45-70. 

OgH^Ag^NgOy requires Ag = 45-95 per cent. 

Digitized by 



This analysis shows that, in the formation of this salt, the hydrogen 
of the hjdroxyl group, as well as that of the carhoxyl group, has 
been replaced by silver (see p. 176). This silver salt is moderately 
easily soluble in warm water, and may be recrystallised from this 
solvent, but, apparently, with a good deal of decomposition. 

The faintly alkaline, dilute, orange-coloured solution of the am- 
monium salt of dinitrohydrozyparaxylic acid gives no precipitate with 
copper aeekUef barium chloride, or zinc acektte, but lead acetate produces 
a very sparingly soluble, heavy, yellow precipitate, and ferric chloride 
a brownish-red precipitate. 

Dibromhydroxi/paraxylio Acid, ^ 

The action of bromine on hydroxyparaxylic acid was investigated in 
the following way : 3*9608 grams of the pure, finely powdered acid 
was exposed in a glass dish to the action of bromine vapour for 12 
hours, during which time much hydrogen bromide was produced ; after 
removing the excess of bromine by exposure over solid potash in a 
vacuum desiccator, it was found that 3*6932 grams of bromine had 
been absorbed. Eepresenting the reaction by the equation, 

Cj^io^g + 2Br2 = Q^^Brfi^+ 2HBr, 

the increase in weight should have been 3*769 grams. 

A determination of the bromine in this crude substance gave the 
following result. 

0-2378 gram gave 0*2698 AgBr. Br » 48*27. 

CgHgBrjOj, requires Br* 49*38 per cent. 

It is difficult to crystallise the crude substance, and the only way 
this could be accomplished was to dissolve it in a little methylic alcohol, 
add a good deal of chloroform, and evaporate to a small bulk to remove 
most of the methylic alcohol ; if, now, more chloroform is added and, 
after rapidly concentrating, the liquid is allowed to stand, the super- 
saturated solution rapidly deposits well-shaped, prismatic crystals, 
which, on analysis, gave the following results. 

0-1836 gave 0-2246 COj and 0*0428 H,0. C = 33*37 ; H = 2*58. 
0-3218 „ 0-3726 AgBr. Br = 49 26. 
C^HgBrjOg requires = 33*33 ; H = 2*47 ; Br = 49*38 per cent. 

IHbromhydroxyparaxylic acid melts at 204 — 205^ and is very readily 
Bolnble in alcohol, but only sparingly in chloroform ; when pure, it 
crystallifles well from dilute alcohol. 

Digitized by @^Ogle 

192 perkin: oxidation of sulphocamphylic acid. 

It is a very stable substance, and may be boiled for some time with 
caustic potash without decomposing. 

A determination of the molecular weight of this bromo-acid by 
Raoult's method^ using acetic acid as the solvent, gave M = 333| 
whereas the value calculated from the formula CgHgBrjOg is 324. 




This zylenol is produced, as explained in the introduction, when the 
calcium salt of hydroxyparaxylic acid is distilled with lime. 

The calcium salt was prepared by boiling finely divided hydroxy- 
paraxylic acid (30 grams) with water and a slight excess of pure 
calcium carbonate, filtering, and evaporating nearly to dryness, when, 
on standing, the calcium salt separated in warty masses. 

The dried salt was finely powdered, mixed intimately with slaked 
lime (50 grams), and distilled in small quantities from small retorts 
over a blow-pipe, as great heat is necessary to bring about the de- 
composition. The distillate, which was an aqueous liquid containing 
dark-coloured, semi-solid masses and smelling very strongly of phenol, 
was mixed with a slight excess of caustic soda, decolorised by boiling 
with animal charcoal, filtered, and the filtrate acidified with hydro- 
chloric acid. The oily precipitate solidified rapidly and almost com- 
pletely, and after collecting, washing, and spreading on a porous plate, 
an almost colourless, crystalline mass (3 grams) was obtained. On 
submitting this to distillation, nearly the whole passed over at 218^ 
(760 mm.) as a colourless oil which, on cooling, at once solidified. The 
distillate, dissolved in a considerable quantity of water at 70°, on 
slowly cooling, deposited very long, colourless needles, which were 
dried, first on a porous plate and then at 60% and analysed. 

0-1415 gave 0-4075 OOj and 0-1053 H^O. C- 78-54 ; H=.8-27. 
CgH3(CH8)2-OH requires 0= 78*69 ; H = 8-19 per cent. 

This substance, which has been found to be 1 : 2'Xylenol{Q), melts at 
74—75°, and is moderately soluble in hot water; its cold, saturated, 
aqueous solution gives a distinct, although not intense, bluish-violet 
coloration with ferric chloride, and, on standing, the solution rapidly 
becomes opaque (compare Tohl, Ber,, 1885, 18, 2562; Nolting and Forel, 
ibid., p. 2673). 

Tribromo-xylenol, C^Br3(CH3)j*OH. — ^This substance, which has 
already been described by Tohl {loc, cit.), was prepared by adding about 
0*5 gram of the xylenol obtained in the above experiment, in small 

Digitized by VjOOQIC 


quantities at a time, to 3 grams of bromine, care being taken to 
moderate the vigorous reaction by cooling; after 2 hours, the 
product was poured on to a glass dish, 'and left over potash in a 
desiccator until the excess of bromine had been removed, the residue 
being purified by recrystallisation from the dilute alcohol. 

0-2840 gave 04440 AgBr. Br. = 66-62. 

CgHyBrgO requires Br = 66-86 per cent. 

This tribromo-zylenol crystallises from dilute alcohol in long, pale- 
yellow needles which melt at 184^, and it thus agrees in its properties 
with the tribromo-derivative which Tohl obtained by brominating 

MMMyparaxylio Acid, [(CHg)^ : OOOH : OCHg -1:2:4:6]. 

With the object of obtaining further information as to the constitu- 
tion of the acid CgHj^Og, now known to be hydrozyparazylic acid, it 
was decided to endeavour to oxidise one or more methyl groups, and 
in order to do this it was necessary, in the first instance, to protect the 
hydroxyl group in the usual way, namely, by substituting methyl or ethyl 
for hydrogen. For this purpose, in the first experiments, the methylic 
salt of the acid (16 grams) was mixed with a solution of sodium 
(2 grams) in methylic alcohol and methylic iodide (26 grams), and the 
mixture heated first at 96 — 110° for 2 hours, and afterwards at 
120 — 130° for 2 hours. The light-brown product was evaporated 
on the water-bath, water added, and the heavy oil extracted with 
ether ; the ethereal solution, on evaporation, deposited an oil which did 
not show any signs of crystallising, and which, doubtless, consisted of 
the methylic salt of methoxyparaxylic acid,OOH:3-OgH2(CH:3)2'OOOCH3. 
This was not analysed, but at once hydrolysed by boiling with excess 
of methyl alcoholic potash for 2 hours ; after evaporating the methylic 
alcohol, and acidifying the aqueous solution of the residue, a very 
bulky, white precipitate was obtained, which was purified by re- 
crystallisation from 60 per cent, alcohol. 

0-1376 gave 0-3366 00^ and 0-0823 HgO. C = 66-66 ; H = 664. 
CgH2(CH3)2(OCH3)-COOH requires = 66-66 ; H-6-66 per cent. 

Methoxyparaxylic acid melts at 170 — 171°, and distils at a high 
temperature, apparently without decomposition. It is very sparingly 
soluble in hot water, and almost insoluble in cold water ; it crystallises 
from boiling water in woolly masses consisting of fine needles. It is 
readily soluble in acetic acid, methylic alcohol, chloroform, and ethylic 
acetate, but only sparingly in carbon bisulphide and light petroleum ; it 
crystallises splendidly from ethylic acetate in long, colourless prisms. 

Digitized by VjOOQIC 


Ethylie Ethoxupa/raxylate, O0fi^<^^^{GR^^*QOO0^^. 

This was prepared by heating ethylie hydroxyparaxylate (2 grams) 
with sodium (0*26 gram) dissolved in alcohol, and ethylie iodide 
(5 grams) at 120 — 130° for 2 hours, evaporating the product on the 
water-bath, and adding water, when an oil separated which rapidly 
solidified. The crystalline mass was collected, washed with water, left 
in contact with porous porcelain until free from oil, and recrystallised 
from dilute methylic alcohol, when glistening prisms were obtained 
which gave the following results on analysis. 

0'1496 gave 03841 COg and 0-1110 HgO. C- 70-02 ; H=.8'24. 
CijHigOg requires C = 70-27 ; H = 8-11 per cent. 

Ethylie ethozyparaxylate melts at 50 — 5P, and is readily soluble in 
methylic alcohol, light petroleum, benzene, ohloroform, and carbon 

Ethoxypa/raxylic Acid, OG^K^^OoK^iGlI^^^COOK. 

When boiled with alcoholic potash, ethylie ethozyparaxylate is 
readily hydrolysed, and if the product is mixed with water and evapo- 
rated until free from alcohol, the addition of hydrochloric acid to the 
cold aqueous solution precipitates the acid as a white powder which 
crystallises from methylic alcohol in beautiful, glistening, prismatio 

0-1476 gave 03656 COg and 0-0954 H^O. O = 67-55 ; H ^ 7-18. 
0-1332 „ 0-3311 OOg „ 0-0867 HjO. = 67-79 ; H«: 7-21. 
CjiHiPg requires C« 68-05 ; H = 7-22 per cent. 

Ethoxyparaxylic acid melts at 173 — 174°; it is sparingly soluble in 
water and light petroleum^ but readily in ethylie and methylic alcohols, 
benzene, and ethylie acetate. 

MethoxymeUiylterejMhalic Acid, ^^fPf Y'^s 


Methoxyparaxylic acid is moderately readily acted on by potassium 
permanganate in alkaline solution, with formation, in the first instance, 
of methoxymethylterephthalic acid. In studying this oxidation, the 
pure acid was dissolved in dilute sodium carbonate, heated to boiling, 
and a strong, hot solution of potassium permanganate added drop by 
drop until the colour remained permanent for some minutes; the 

Digitized by 



excess of permanganate was then destroyed by the addition of a drop 
of alcohol, and,, after filtration, the* filtrate evaporated to a small bulk 
was acidified with hydrochloric acid. 

The bulky, white precipitate which separated at once, was collected, 
washed well, and crystallised from dilute alcohol. 

0-1263 gave 02648 CO^ and 0-0570 H^O. C = 57-18 ; H = 5-01. 
01643 „ 0-3223 CX)j „ 0-0679 H^O. = 56-95; H = 4-88. 
CH30-CeH,(CH8)(COOH)2 requires = 57-14; H = 4-76 per cent. 

Methozymethylterephthalic acid, when heated in a capillary tube, 
sinters at about 250°, and melts completely at 267°. It is only 
sparingly soluble in cold water, but crystallises well from large 
quantities of boiling water. With ferric chloride, the aqueous solution 
gives an ochreous precipitate. 

HydroxyTMihylterejihihcilic Add, 

In order to prepare this acid, methozymethylterephthalic acid 
(3 grams) was boiled with fuming hydriodic acid (30 c.c.) in a reflux 
apparatus for 2 hours. The liquid, which, on cooling, became filled 
with crystals, was diluted with water, filtered, and the crystals 
washed with a little dilute sulphurous acid to remove iodine. In this 
condition, the crystals are nearly insoluble in water, but if the solution 
in dilute alkali is heated to boiling and acidified, the acid does not 
separate at once, and the filtered solution, on slowly cooling, deposits 
a voluminous precipitate consisting of microscopic needles, The pre- 
cipitate was collected, washed well, and analysed, with, the following 

0-1512 gave 0-3063 COj and 00580 HjO. = 55-24 ; H =- 4-26. 
OH- CeHj(OHg)(0OOH)2 requires = 55-10. H = 408 per cent. 

Hydroxymethylterephthalic acid melts at 280 — 283°, and its solu- 
tion in methylic alcohol gives, with ferric chloride, an intense reddish- 
violet coloration. 



Digitized by VjOOQIC 

196 ackroyd: researches on moorland waters. 

XXIII. — Researches on Momiand Waters. L Acidity. 

By William Aokeotd, F.LO. 

Moorland waters are consumed at present by over five and a quarter 
millions of people in Yorkshire, Lancashire, Westmoreland, and 
Cumberland. Their well known plumbo-solvent action has been con- 
sidered of such importance as to warrant a €k}yernment investigation, 
the results of which will be found collated by Mr. W. H. Power, F.K.S., 
in the Supplement to the Twenty-third Annual Beport of the Local 
Government Board, pp. 332 — 124, and it is there demonstrated from 
the work of Mr. and Mrs. Atkinson and Drs. Barry and A. C, 
Houston that acidity and lead dissolving power run parallel to each 
other. In common with many analysts, I have long associated 
plumbo-solvent power with acidity, and in an investigation made for 
the Halifax Corporation concerning Moorland waters from many 
gathering grounds, I reported to this effect in 1895. References to 
the question are made in papers by Messrs. A. H. Allen,C. Rawson, and 
T. Whitaker (Journal qf tkt Society qfJDyw and, Colourieia, 5, 64—71). 
The question of acidity is undoubtedly of primary importance in con- 
sidering this class of waters, and the following experiments and 
observations thereon will, therefore, be of interest. 

Method of Estimating Acidity. 

The solutions to be dealt with may be under or a little over xxrornr^^ 
of normal, and as concentration by the usual methods is objectionable, 
I have adopted the following course of procedure as being both the 
easiest and quickest, and the least likely to introduce errors. 

To 100 C.C. of the water in a beaker, 1 or 2 drops of alcoholic 
solution of phenolphthalein are added, and N/100 alkali is run in 
from a burette until a very slight pink tint is obtained ; this gives the 
total acidity. Through a little more than 100 a c. of water, air, free from 
carbonic anhydride, is rapidly aspirated for half an hour ; titration of 
100 c.c. of the water thus treated gives the residual or organic acidity, 
the difference between the first and second titrations being the volatile 
acidity, and as this is very largely, if not entirely, due to carbonic acid, 
it may be termed inorganic acidity. As alcoholic solution of phenol- 
phthalein, when used as an indicator, is sometimes neutralised by the 
addition of a little alkali, it is well to mention here the precaution that 
not more than 1 or 2 drops of such a solution should be used in titrating 
moorland waters, or somewhat discordant results may be obtained ; 
this effect I have traced to the acetate present in the indicator solvent. 

Digitized by VjOOQIC 




In my earliest observations of waters taken from places oontiguons 
to moorlands, there was gradual decrease of acidity on keeping ; this 
I was at first at a loss to account for, but it was eventually put down 
to diffusion ; the Winchester quarts containing the waters were kept 
in a room where they were freely exposed to sunlight, and they were 
only opened as required for purposes of analysis. In three of these 
samples, A, B, and C, the following percentage changes were observed. 


Day of testing. 


























A was from a private well not far removed from moorland ; B was 
from another well of the same kind, and C was from a trench 7 feet deep 
dug into the side of a moor-oapped hill of Millstone Grit.* 

It was these results which suggested the advisability of rapid 
aspiration of air free from carbonic anhydride as a necessary part of any 
method of estimating acidity. As a preliminary trial, distilled water 
was charged with carbonic anhydride ; 100 c.c. took 60*1 c.c. of N/100 
alkali, but after aspiration for half an hour only 0*3 c.c. of alkali was 
required for neutralisation, which was the initial acidity of the distilled 

* Some other analjrtical data concerning these samples are as follows. 
Parts per 100,000. 



Free NH,, 


N as nitrites 



c.c. of 

N/100 alkali 

required per 

100 C.C. 












It may also be mentioned that all the waters referred to in this paper are of 
permanent hardness only. 

Digitized by VjOOQIC 



water. A sample of the water C was now sent for and submitted to 
rapid aspiration for 30 minutes. Its acidity was reduced to 26 per cent, 
of the original. 

Water taken direct from moorland catchwater or stream maj 
exhibit little alteration on aspiration. Thus a water from a clough 
receiving its supply from the adjacent moorland showed no appreci- 
able alteration by diffusion in 15 days, neither was the effect of 
aspiration any more marked. Some freshly-collected samples from 
very peaty gathering grounds gave respectively 14, 11, 5, and 6 per 
cent, of volatile acidity, calling the total acidity 100. The following 
figures were obtained with waters from moorland reservoirs. 

cc, qf JV/100 alkali required per 100 cc. 






E. .. 

Centre of large peaty reaeiroir 
Small reseryoir 



1 '2 =22 per cent. 
0*6=88 ,. .. 


»» i> 

0-9 = 81 „ „ 

Configuration of Gathering Grounds. 

Some of my observations appear to throw a new light on the 
somewhat old problem, why one gathering ground should yield waters 
which acted much more quickly on lead, or, in other words, were so much 
more acid, than others. This has been a fruitful subject of discussion 
and of municipal inquiry, in the course of which expressions have been 
evolved, and have come into use as synonymous with bad and good 
waters as, for instance, **high " and " low-level " supplies, the former 
being regarded as productive of epidemics of plumbism. I have found 
that mere altitude of gathering grounds is not sufficient to account for 
some of the striking differences observed in these waters. Certainly, 
one of the worst waters with which I have had to deal is a high-level 
one, the reservoir being 1300 feet above sea-level, and better waters 
have been had from reservoirs 860 to 1020 feet high ; on investi- 
gation, however, it appeared that some of the latter have gathering 
grounds rising to a higher level than in the former case. The 
** high " and ** low-level " distinction has, therefore, seemed to me of 
little or no value, and I have come to the conclusion that, other 
things being equal, it is rather a question of the average gradient of a 
gathering ground. The reasons for this conclusion are that the 
acidity lis due to the peat through which the water passes, and that 
there will be less time for solution or the taking up of the acid wher^ 

Digitized by VjOOQIC 

I. AciDiry. 


the gradient is a high, one than where it ifl a low one. I have 
submitted this hypothesis to the test in a comparison of four 
gathering grounds on like geological strata — ^the Millstone Grit — with 
the foUowing results. 

Gathering ground. 

Reservoir or 

^'^\eYdr*"* ^"^*^°*- 

Total acidity. 


1880 to 1460 feet 1 in 44 



980 „ 1100 „ 1 in 28 to 1 in 28 



986 „ 1480 „ 1 1 in 12 
1076 „ 1344 „ 1 in 12 




Corrdation qf Acidity cmd Absorbed Oxygen Nwmbere. 

Another fruitful source of speculation has been the nature of the 
acidity of moorland waters. The earliest idea attributed it to 
mineral acids; now it is put down to peaty acids or, to be more precise, 
to humic acid. It appeared not unlikely that moist combustion com- 
bined with organic acidity determinations might throw further light 
on the subject, as, for example, some direct relation between the 
absorbed oxygen numbers and those for organic acidity, on the assump- 
tion that the organic acidity represents the presence of a carbon com- 
pound of definite composition. Forschammer's permanganate process 
was employed, the length of the time allowed being 4 hours. The 
following observations are for waters from widely different moorland 

Parts per 100,000. 






From a large serrice reservoir of public 










AnoUier supply of inferior quality.... 
From the feeder to a reservoir of com- 
pensation water; running 



From the same source as'k ; feeder 
f^t^gnant ....x. 


Mountain stream 

1-4 - 

By inter- 

1 0126 

Digitized by VjOOQIC 



The absorbed oxygen is here converted into organic carbon by using 
Sir E. Frankland's factor 2*4. This step was checked by referring 
back to his analysis of a moorland water, made for official purposes, 
and of which I had a duplicate sample, 0. The organic acidity was 
1*4 ; this gives 0*125 part of organic carbon per 100,000 by interpo- 
lation, as against 0*15 part returned in the analysis. 

The direct relation looked for only appears in the first two observa- 
tions. Afterwards, as the acidity increases and the yellow colour 
deepens, the carbon appears to increase abnormally, as shown in curve 
I. in the diagram. 

M ,*« >ff *» JM 4« 90 9^ »t. a^ M «•• iM 41* W *'• ^ 

CAf^BON Parts PcR 100000 

The general assumption that the acidity is due to humio acid is 
somewhat wanting in precision, seeing that some half-dozen humic acids 
have been found oj late years. The basicity of two have been ascer- 
tained, namely, Detmer's octobasic acid, C^Q'H^fi^ {W(Ut8' Dictumaryf 
7, 648), and Berthelot and Andre's tribasic acid, C^^K^fi^ (Abstr., 
1891, 1089). The curve for the carbon of the latter approaches 
nearest to my observations for moorland waters (curve II, diagram) ; 
yet it is so far away from it that one is warranted in concluding 
that the organic acidity of moorland waters is due to an acid or 
acids of lower equivalent than that of the humic acids whose basicity 
has been ascertained up to now. 

The Bobouoh Labobatobt, 

Halifax, Yorkshire. 

Digitized by VjOOQIC 


XXIV.— 7%6 Nutrition of Yeast. Part I. 

By Arthur L. Stern, D.Sc. 

Solutions in which yeast grows freely and ferments yigorously, 
always contain, besides a fermentable sugar, certain inorganic salts and 
nitrogenous organic compounds required for the growth and development 
of the yeast. 

The composition of yeast ash indicates that the inorganic constituents 
are chiefly potassium and magnesium phosphates ; small quantities 
of calcium, silicon, sulphur, and iron are also usually present. 
A. Mayer (UtUersttohufigen iiber die alko?iolische GHhrung, Heidelberg, 
1869), who examined the nutritive value of a large number of salts, 
concluded that magnesium and potassium phosphates could supply all 
the inorganic nutriment required by yeast. There is, however, con- 
siderable conflict of evidence amongst later investigators as to the 
necessity for any other inorganic food. Lintner (Lehrbtush der 
Bierhraner, p. 427) says sodium and iron are not necessary ; H. Molisch 
{Bet. Cenir^ 1894, 167) says iron is necessary, but that calcium is not ; 
Wehmer {Beitrage zur Kentnisa einheimischer Filze, p. 158) disputes the 
necessity for iron ; whilst Leyffert (Zeitschr, /v/r das gee, Brau,^ 1896, 
318) states that yeast degenerates in the absence of calcium salts. 

The necessity for sulphur in some form or other is generally assumed, 
although it does not appear to have been ever proved. 

Mayer (loc, eit) also examined the nutritive value of a number of 
nitrogenous substances, and found that the most suitable were aspara- 
gine, aUantoin, carbamide, and diffusible peptones ; ammonium salts 
were not very useful ; colloidal albuminoids, like white of egg, fibrin, 
and casein, were useless, as also were nitrates. ' These deductions are 
in general agreement with those of later observers, amongst whom 
may be mentioned Matthews {Jour, Inat, Brewing^ 1897, 369), Ehrich 
{Der Bierbrauer, 1895, 145, 161, and 177), and Wahl and Hanke 
{Amer. Brewers* ^ev., 7, 32); the latter state that yeast exerts a 
selective influence on the nitrogenous constituent-s of malt worts, 
assimilating most of the amides, and only small portions of the pep- 
tones and albuminoids. 

Whilst a fairly general agreement prevails that certain substances 
aie necessary for the growth and development of yeast, yet very little 
is known as to the amount required, or of the effect which a variation 
of the amount and proportion of these substances produces on the 
quantity, composition, and properties of the yeast. 

Almost the only work in this direction is that of Hayduck {Zeit- 
sehriftfwr SpirUtisind., 1881, 173), who determined the effect of vary- 

Digitized by 



ing the amount of nitrogenous nutriment on the weight, the amount 
of nitrogen, and the fermentative power of the yeast. He found that 
the amount of nitrogen in yeast can vary very considerably, that any 
increase of nitrogenous nutriment, in the form of asparagine, up to 
0*05 gram nitrogen per 100 c.c, results, not only in an increase in 
the weight of yeast, but also in an increase in the percentage of 
nitrogen present in it, but any increase in the nitrogenous nutriment 
above that mentioned neither increases the weight of yeast nor the 
percentage of nitrogen in it. 

Hayduck determined the fermentative power of the yeasts obtained 
in these experiments, and found that those which contained the 
highest percentage of nitrogen had the greatest fermentative power. 

Kusserow (Brermerei £^eitung, 1897, 317) states that the addition of 
asparagine to a fermenting solution increases the rapidity of fermenta- 
tion, but diminishes the yield of yeast ; Heinzlemann (Zeiisehr, fur 
Spiritu8tnd,f 1897, 296 and 311) makes a similar statement with 
regard to the rapidity of fermentation. 

The effect of varying the amount of inorganic nutriment has never 
been exhaustively studied, and the statements made on the point are 
not concordant; Kusserow, for instance {loe, dt), found that an 
increase of inorganic nutriment increased the rate of fermentation, 
the yield of yeast, and its fermentative power, whereas Heinzlemann 
{loc, cit,) states that it produces no effect on the rate of fermentation. 
A. G. Salomon and W. de Y. Matthews {Joum, Soc. Chem, Ind,, 1889, 
376) found that the addition of phosphates to malt worts decreased 
the assimilation of nitrogen and phosphorus by the yeast. 

The experiments described in this communication were undertaken 
with the object of determining the effect which the variation of the 
amount of inorganic and nitrogenous nutriment produces on the crop 
of yeast, the assimilation of nitrogen, and the fermentation of the 
sugar in a given time. 

Although some of these problems have been solved by Hayduck, 
experiments of this nature, to be strictly comparable, must be carried 
out under absolutely identical conditions; moreover, they acquire 
additional value when considered in relation to similar work carried 
out under somewhat different conditions. On this account, it was 
considered advisable to make the investigation as complete as possible. 

Method qf £aBperimerU, 

The experiments were made in flasks holding rather more than a 
litre, into which the required amount of mineral and nitrogenous 
nutriment was introduced, together with 500 c.c. of a 10 per cent, 
solution of pure dextrose, and the neck of the flask was plugged with 

Digitized by 



cotton wool. The solution was then boiled for some minutes in order 
to sterilise the contents, and when cold the yeast was added. The flasks 
were kept at a temperature of 24°, and well shaken twice a day, when 
the odour of the evolved gases was noted. 

At the end of the stated time, the yeast was filtered off, washed 
tmoe with cold water, dried at 100°, and the nitrogen in it determined 
by Kjeldahl's method. The fermented solution was boiled until the 
alcohol was expelled, cooled, made up to its original volume, and its 
optical activity observed, from which the amount of dextrose remain- 
ing in solution could be calculated. 

The yeast employed throughout this investigation was prepared 
from a Burton pitching yeast. An Esmarch roll culture was made 
with wort gelatin, and when, after incubating, the colonies appeared, 
one was introduced into a Pasteur flask containing sterile hopped 
wort. Three such flasks were thus seeded, and from these all the 
yeast used in this investigation was obtained. Neither in their micro- 
scopic appearance nor in the flavour of the beer produced by them 
eould any difference in these yeasts be detected, and they in no 
appreciable way differed from normal Burton yeast. 

In order to obtain sufiScient yeast for each series of experiments, 
the fermented wort was decanted from one of the Pasteur flasks, which 
was then filled up with sterile hopped wort, and well mixed with the yeast 
sediment ; a portion was then poured into a flask containing 500 c.c. 
of sterilised hopped wort. Fermentation soon began, and when 
finished the yeast fell to the bottom ; the fermented wort was then 
poured off, and the yeast well mixed with sterilised water. The 
number of yeast cells in a measured volume of this was counted, and 
sufilicient added to each flask containing the solution to be experimented 
on, to give 1*5 cells per 1/4000 cubic mm. 

Asparagine was chosen as the nitrogenous nutriment, since it is 
easily obtained in a pure state, and previous observers have found it 
a very suitable yeast nutriment. It contains 1 8 '7 per cent, of nitrogen. 
The mineral nutriment A was prepared by dissolving the ash of 
yeast in a weak aqueous solution of tartaric acid ; B was prepared by 
dissolving potassium phosphate, magnesium sulphate, and calcium 
sulphate in water. 

A. B. 

PaUanam silicate 0'8 Calcium sulphate 8'2 

Caldnm hydrogen phosphate I'd Magnesium sulphate 15*9 

Magnesium hydrogen phosphate. . 1 '5 Potassium hydrogen phosphate . . . 80 *9 
Potassium hydrogen phosphate... 87 '9 

100-0 1000 

Digitized by VjOOQIC 



Four series of experiments were made with nutriment A. In the 
first of these, 0*058 'gram per 100 c.c. of A was used, and varying 
amounts of asparagine were added in different experiments. After 
7 days, only about half of the sugar was fermented. 

In the second series, the conditions were the same, except that the 
fermentations were continued for 12 days ; here, about three-quarters 
of the sugar were fermented. 

In the third series of experiments, the amount of inorganic nutri- 
ment was doubled, when as much sugar was fermented in 7 days as 
was fermented in the second series in 12 days. 

In the fourth series, three times the original amount of inorganic 
nutriment was used, but this had no further effect. 

The presence of a considerable amount of unfermented sugar indi- 
cates the absence of some essential yeast nutriment. 

A series of experiments were then made with inorganic nutriment 
B, almost identical with that used by Hayduck, and chiefly differing 
from A by the presence of sulphates. In these experiments, the sugar 
was almost completely fermented in 7 days, whilst the amount of yeast 
and of assimilated nitrogen was much greater than in the first series 
of experiments. 

As all other conditions were identical, this must be due to the 
difference in composition between nutriments A and B, and in order 
to determine if the presence of sulphates in the one and their absence 
from the other was the cause of the difference in their behaviour, 
experiments were made with nutriment A with the addition of 
sulphate, and with nutriment B with its sulphate replaced by chloride. 
These experiments proved that the absence of sulphate was the cause 
of the difference observed. This question of sulphur nutriment is of 
considerable interest, and will be treated fully in a future paper. 

Nutriment B was used throughout the remainder of the experiments. 

From comparative experiments, it was found that the presence of 
iron has no effect on any of the functions of yeast investigated in this 
paper, and cannot be said to be essential to the growth and develop- 
ment of the yeast. 

When all the elements necessary for the growth of the yeast were 
present, it was found that the greatest amount of nitrogen assimilated 
by the yeast (0*3246 gram) in any experiment was 0*0227 gram per 
100 C.C., and, although in some experiments more than six times this 
amount of nitrogen was supplied, yet the yeast was never able to assimi* 
late more than the amount just mentioned. From experiments made by 
Nageli and Low {Zeitaehr./ur das gesammte Brauioesen, 1878, 337) and 

Digitized by VjOOQIC 


otherSy it has been found that the weight of the inorganic constituents 
of yeast is approzimatelj equal to the weight of the nitrogen it con- 
tains. In round numbers^ then, a supply of 0*026 gram per 100 c.c. 
of inorganic nutriment and of nitrogen as asparagine is as much as 
the yeast (0*325 gram) is able to assimilate under the conditions of 
these experiments. This may be called a normal supply. 

A. — Effect qf Varying the Amotmt qflfitrogenoiu Ifutriment an: 

1. The Amount of NUrogen Asaimilctted. — When the normal or any 
larger amount of inorganic nutriment is present, any increase of 
nitrogenous nutriment above the normal has very little effect on the 
amount of nitrogen assimilated. The experiments point rather to a 
decrease of this as the amount of nitrogenous nutriment is increased. 
If the amount of nitrogenous nutriment is below the normal, the 
assimilation of nitrogen is much less (Expts. 1, 5, and 6 ; see p. 207). 

2. The Percentage qf NUrogen AesvmUated. — ^This is greatest when 
the supply of nitrogen is least. 

3. The Percentage of Nitrogen contained in the Teaet. — ^This does not 
vary when the nitrogen supply is normal or greater than normal ; 
when, however, the nitrogen supply is below normal, the percentage 
of nitrogen contained in the yeast is low (Expts. 1, 5, and 6). 

4. The Percentage qf Sugar remaining Utifermented. — ^This does not 
materially vary, even when the supply of nitrogen is below the normal ; 
if, however, the inorganic nutriment is present in only normal amount, 
the nitrogen supply being below the normal (Expt. 1), the amount of 
sugar remaining unfermented is large. 

5. The Weight qf the Teaet Crop.— This is practically unaffected 
except in the case just noted (Expt. 1), when the yeast crop is small. 

B. — Effect of Varying the Amount qf Inorganic Nutriment on: 

1. TheAmowtU of Nitrogen Asevm/Uated, — When the nitrogen supply 
is below normal (Expts. 1, 5, and 6, p. 207), an increase of the inorganic 
nutriment has little effect on the assimilation of nitrogen. When the 
nitrogen supply is normal (Expts. 2, 7, and 8), an increase of the 
inorganic nutriment causes an increase in the nitrogen assimilation. 
When the nitrogen supply is above the normal (Expts. 3, 4, 9 — 23), 
little eff eot is produced by increasing the inorganic nutriment, although 
the trend of the figures indicates a slight increase of nitrogen assimila- 
tion as the inorganic nutriment is increased. 

3. Hie Percentage qf Nitrogen contained in the Teaet. — ^When the 
nitrogenous nutriment is below the normal (Expts. 1, 5, and 6), an 

^o"* ^"^- Digitized by C?oogle 


increase of the inorganic nutriment causes a decrease in the percentage 
of nitrogen contained in the yeast. When the nitrogenous nutriment 
is normal or greater, an increase of the inorganic nutriment has 
practicaUy no effect on the percentage of nitrogen in the jeast. 

4. The PeronUage qf Sugtx/r rtmaining UnfwmBnUd, — ^When the 
supply of nitrogen is below the normal (Expts. 1, 5, and 6), an increase 
of the inorganic nutriment causes a decrease in the amount of sugar 
remaining unfermented. When the nitrogen supply is normal or 
greater, an increase of the inorganic nutriment has little effect on the 
amount of sugar fermented, although the trend of the figures indicates 
that this causes a decrease in the amount of sugar remaining 

5. Th6 Weight qf the Teast Crop. — ^This is increased a little by 
increasing the amount of inorganic nutriment up to 0*25 gram, the 
effect being most marked when the nitrogen supply was least. 

The results of the experiments in which one of the essential 
elements (sulphur) of yeast food is absent, present much the same 
genwal features, but the effect of increasing the inorganic nutriment 
is much more marked. 

So far as might be expected considering the difference in the 
variety of the yeast used and the differences in the conditions, 
these experiments are in general agreement with those of Hay- 
duck. It is, however, noteworthy that, whilst Hayduck found that 
the fermentative power of yeast was proportional to its nitrogen 
contents, yet in experiments 5 and 6, in which the proportion of 
nitrogen in the yeast was very low, the fermentation proceeded just 
as far as in the other experiments. 

The most striking result of this investigation may be stated as an 
enlargement of a proposition which Hayduck deduced from his more 
restricted work. 

Any increase of nutriment beyond a definite limit will not materially 
increase either the amount of nitrogen assimilated by the yeast, the 
percentage of nitrogen or the weight of the yeast, or the amount of 
sugar fermented. This limit is but little greater than the largest 
amount which the yeast is able to assimilate under the conditions of 
the experiment. 

October, 1898. 

Digitized by VjOOQIC 




^ i 

10 OQOqO 
K» -« 10 10 




« ** 

§•3 II. 




>4 ;^ B o 





> oop 


e« o oo <o 
p pop 


9 *i 




A 00 <D 00 



^ 5 









00 00 






€4 O 00 
r-l .09 «»0 
p •p -p 


• •» -O "^rH * * 

O -O "O 

00 o» 

o o 

a 5 s 5 



OCfl "* 







-3 - 

S 5 2 


•^i rH 00 rl 00 

rHC^<« 10 CO 







oc»r^ Oftoo 






00000 1 




•»r^ •» 







ko ^ 

.^-^ •« 


"»-l •> 



?s ' 


1-1 1-4 

MC4 01 











Digitized by 



XXV. — An Isomeride of Ama/rine. 

By H. Llotd Snape^ D.Sc.y Ph.D.^ and Arthur Bbooks, PhD. 

AoooBDiNO to Laurent (Armaim, 1844, 62, 366), benzoylazotide, on 
distillation, yields a mixture of ** amarone " and lophine, and we have 
shown (Trans., 1897, 71, 628) that amarone is tetraphenylazine ; a 
further examination of the products of distillation has shown that a 
third base, C^iH^g^s' ^ ^^ present; this melts at 198°, and when 
subjected to dry distillation yields lophine, Cj^Hj^N,. Like amarine, 
with which it is isomeric, it forms salts with one equivalent of add, 
and yields a mono-silver derivative, which, on heating, decomposes, 
with formation of lophine ; the silver derivative combines directly with 
ethylic iodide, hence the base contains an imido-group. 

The new base differs from amarine in that it has a much higher 
melting point; moreover, its chromate does not undergo oxidation 
when boiled with glacial acetic acid, whereas that of amarine chromate, 
under similar conditions, yields lophine. 

Neither is this base identical with the iso-amarine (melting at 176°) 
described by Feist and Arnstein (Ber.y 1895, 28, 3177), nor with the 
isomeride of amarine (melting at 269°) mentioned by 0. Fischer 
{Annalen, 1888, 246, 288) as having been prepared by Lorenz Kohler. 

The isolation of the base and the preparation of some of its deriva- 
tives are described below. 

Isolation of the New Base. 

1. From the Products of DistUUUion qf Benzoylazotide, — ^The higher 
distillation products of benzoylazotide, or preferably the residue which 
is left in the retort when that compound is heated to about 216°, were 
treated with excess of warm, dilute hydrochloric acid, which diAsolved 
the base and some resinous matter, leaving the tetraphenylazine and 
lophine undissolved ; the base was then precipitated with ammonia, 
and purified from resinous matter by suspending it in a large quantity 
of water, adding hydrochloric acid in quantity insufficient to dissolve 
the whole of the base, and heating the solution by passing in steam ; 
the base was then reprecipitated by ammonia. By repeating this 
process several times, the resinous matter, which is insoluble in pure 
water, is gradually removed, as it remains adhering to the walls of 
the vessel. 

2. From Mather Liquors. — ^The alcoholic mother liquors obtained in 
the preparation of benzoylazotide from benzaldehyde and ammonium 
cyanide (Trans., 18971,^71, 629), especially when excess of the aldehyde 

Digitized by VjOOQIC 


was employed^ were also found to contain the new base ; the volatUe 
sabstanoes were removed from these liquors by heating with steam, 
and the dried and powdered residue was extracted with small quanti* 
ties of alcohol and ether. A crystalline substance melting at 267° was 
left behind, which, on analysis, proved to be lophine, whilst the new 
base was extracted by dilute hydrochloric acid from the residue left on 
evaporating the alcoholic ethereal solution. 

It was subsequently found that the most convenient method for ex- 
tracting the base from the benzoylazotide mother liquors consisted in 
acidifying with hydrochloric acid, diluting, and distilling with steam ; 
the hydrochloride of the base was thus obtained in solution, whilst the 
resin, benzoylazotide, and the hydrochloride of lophine remained un- 
dissolved. The last-named salt was isolated in octahedral crystals, in 
one experiment, by crystallisation from chloroform ; but generally the 
residue was subjected to dry distillation at 200 — 215°, and the residue 
left in the distillation flask worked up in accordance with the method 
first given for the isolation of the base. Most of the base used for 
the experiments recorded in this paper was prepared by the method 
just described. 

In all cases, the hydrochloride was precipitated by ammonia, and the 
base recrystallised from alcohol, when it was obtained in colourless 
needles which are soluble in benzene and chloroform, and melt at 198° 
without decomposing. The alcoholic solution does not become lumi- 
nous in the dark after treatment with caustic potash (distinction from 

01778 gave 05500 00, and 00990 H,0. = 8436 ; H = 613. 
01885 „ 15-3 C.C. moist nitrogen at 10'5°and 760-1 mm. N = 9-70. 
OjaHjgNj requires = 84*56 ; H « 6-04 ; N = 940 per cent. 

2^ HydroekUmde, OsiH^gN^HOi + HgO.— A solution of the base in 
warm, dilute hydrochloric acid, decanted from the amorphous salt 
which separated on standing, was evaporated in a vacuum over sul- 
phuric add, when the crystalline salt slowly separated. It has an 
alkaline reaction, is exceedingly soluble in alcohol and chloroform, 
very slightly in ether or benzene, and insoluble in carbon bisulphide. 
It melts at 135 — 140° in a capillary tube, and, on cooling, solidifies to 
a transparent glass. The crystals were dried at 100° until their 
weight was constant. 

0-209 gave 00854 AgCl. 01 = 1005. 

O^HigNyHCl + HjO requires 01 = 1007. 

The salt was not changed by heating with concentrated hydro- 
chloric acid in a sealed tube for 6 hours at 100°. 
PUuinochhrid^y (0,iHi3N,)^H,Pt01e.— This was prepared by 

Digitized by VjOOQIC 


pouring an aqueons solution of the hydrochloride into a solution of 
platinic chloride, the latter being in excess, and the precipitate 
crystallised from rectified spirit. The orange-coloured crystals thus 
obtained begin to lose weight on heating to 220 — 230^, and fuse with 
further decomposition at 230 — 240^. The platinochloride is sparingly 
soluble in hot water, and rather more soluble in boiling alcohol. 

0-2246 gave 00434 Pt. Pt= 19-33. 
0-3741 „ 0-0726 Pt. Pt = 19-40. 

(03iHi8Nj)2,B[2PtClg requires Pt = 19-30 per cent. 

NUrtUe, GgiHjgNs^HNOg + HgO.— The base was dissolved in dilute 
nitric acid on the water-bath, and the crystalline crust which formed 
on the bottom of the vessel, on standing, was redissolved in water ; 
the white crystals which separated from the neutral solution thus 
obtained had an alkaline reaction, and after drying in a vacuum over 
sulphuric acid, melted in a capillary tube at about 165^ 

0-2040 gave 04992 CO^ and 01058 H^O. = 6673 ; H = 5 76. 
0-1807 „ 17-4 ac. moist nitrogen at 15° and 755 mm. N= 11-2. 
02iB[jL8Nj,HN08 + HjO requiresO = 6646 ; H = 5*57; N = 1 1 OSper cent. 

Chramate. — A solution of the hydrochloride, precipitated by potas- 
sium chromate, gave a yellow precipitate which separated from hot 
glacial acetic acid in well-defined, yellow crystals. The salt was not 
altered by heating it with glacial acetic acid for 3 hours at 150° in a 
sealed tube. 

Silver Derivative, OjiHi^NjAg. — Freshly prepared silver oxide was 
dissolved in a dilute alcoholic solution of ammonia, and a solution 
of the base in rectified spirit poured in in a thin stream, agitating 
meanwhile. The almost white compound which at once separated 
was dried at 100° ; after being well washed on the filter with alcohol, 
the light grey powder thus obtained consisted of exceedingly minute, 
transparent crystals melting and decomposing at 210°. It dissolves 
in minute quantity in boiling absolute alcohol, but is insoluble in 
ether, benzene, and light petroleum. When heated with ethylic 
iodide in a sealed tube at 100°, it dissolves, forming a clear solution. 

0-1953 gave 04466 CX), and 00756 HjO. - 6236 ; H - 4-30. 
0-2503 „ 15 o.c. moist nitrogen at 21° and 764-3 mm. N= 6-83. 
0-2041 „ 0-0540 Ag. Ag«26-45. 
OjiHi^NgAgrequireeO = 62-22; H - 4-20; N = 6-91 ; Ag - 26-67per cent. 

Although the relative proportions of base and of ammoniacal silver 
solution were varied in different experiments, nothing but the mono- 
silver derivative could be obtained ; when excess of silver was em- 
ployed, the yield was practically <|uantitativet 

Digitized by VjOOQIC 


ConverHon qf the Silver DerivaUve wUo LojAim. — On heating this 
silver oomponnd to fusion, lophine was formed, but no hydrogen or 
other gas was eyolved. Two grams were heated in a paraffin bath to 
210° ; the transparent, yellow melt, which contained particles of silver 
in sospenaion, solidified to a hard mass on cooling ; the portion ex- 
tracted by hot benzene, after being crystallised from alcohol, deposited 
fine needles sparingly soluble in alcohol, and melting at 267° without 
deoomposition. The alcoholic solution, on adding caustic potash, 
became luminous in the dark when shaken with air. These properties 
indicate that the crystals consisted of lophine. 

Destructive DietiUation qfthe Bcue. 

On heating the base under diminished pressure, decomposition oc- 
curred below 360°, ammonia being evolved and a small quantity of liquid 
distilling over, the last portions of which solidified on cooling ; the 
residue left in the distillation flask and the solid distillate were dis- 
solved in boiling spirit, from which lophine crystallised out in fine 
. needles ; the later crystallisations, which also contained lophine, were 
digested with warm dilute hydrochloric acid, which dissolved the 
unaltered base. On recrystallising the insoluble portion from alcohol, 
more lophine was obtained, and also a small quantity of a colourless 
substance which separated in thin plates and melted at about 120° ; 
as it volatilised below the melting point of lophine, it was easily 
separated from the latter by heating the mixed crystals in an air-bath. 
The crystals of lower melting and boiling point have not been further 

Ukivbbsitt Collbob of Walis, 

XXVI. — Studies of the Terpenes and Allied Compounds. 
Nitrocamphor and its Derivatives. IV. Nitro- 
camiphor as an Example of Dynamic Isomerism. 

By T. Mabtdi Lowby, B.Sc. 

Iv determining the specific rotatory power of nitrocamphor dissolved in 
benzene, it was noticed that the solution showed a considerably greater 
l»vorotatory power when freshly prepared than after keeping, and 
this was found to be a specific property of the substance ; thus a 
specimen which had been crystallised four times showed a change of 
rotatory power from [ a ],> = - 123° to - 104° in aJS per cent, solution in 

Digitized by VjOOQIC 


benzene, whilst after four more crystallisations the limits were - 1 24° and 
—104°. Similar changes of rotatory power are exhibited by ir-bromo- 

nitrocamphor, OgHijBrcCj^ \ which, like nitrocamphor, con- 

tains a secondary nitro-groap; bat no such changes have been 
observed in any compounds from which either the nitro-group or the 
hydrogen atom adjacent to it has been eliminated, that is, the 
phenomenon is limited to seeandary nitro-derivatives of camphor 
Thus no change of rotatory power is observed in freshly prepared 

solutions of the aa'*chloronitrocamphor, CgH^^^jL ^ the aa'-bromo- 
nitrocamphors, CgH^^'OL ^ the salts of pseudonitrocamphor, 

C8Hh<^6 or the anhydride, C8Hh<^6 O^CgH, 

does any change occur in such compounds as a-bromocamphor or 
a-chlorocamphor, from which the nitro-group is absent. The 
phenomenon depends, therefore, on the presence of a particular group 
in the molecule, and must be ascribed to chemical rather than to 
physical changes in the solution ; moreover, as these changes take 
place in all the solvents that have been examined, they must 
obviously be of the nature of isomeric or polymeric change, and are 
independent of any chemical interaction with the solvent. 

It has already been shown that the salts and the anhydride of nitro^ 
camphor are derived from a pseudo-form of the substance to which the 

formula OgHi^^TNO was assigned (Trans., 1898, 73, 997) ; 



these substances have a very high dextrorotatory power, and it is pro- 
bable that pseudonitrocamphor itself would, in like manner, be dextro- 
rotatory. The changes of rotatory power which take place in freshly 
prepared solutions of nitrocamphor can, therefore, be readily explained 
on the assumption that the levorotatory normal form is partially 
converted into the dextrorotatory pseudo-form of the nitro-compound, 
which must then be regarded as isodynamic. In the case of 
ir^bromonitrocamphor, both of the isodynamic forms are known in the 
free state, and, although large differences of rotatory power occur in 
the freshly prepared solutions, ultimately, a condition of equilibrium is 
reached which is the same for both. 

Cawe qf Jfutarotation in the Sugars. — ^The spontaneous changes 
of specific rotatory power which take phice in freshly prepared solu- 
tions of nitrocamphor and ^bromonitrocamphor are very similar to 
those which have been observed in aqueous solutions of many of the 

Digitized by VjOOQIC 


sagaro and of the acids derived from them. The name of birotation 
was originally applied to the phenomenon hj Dubrunfaut (Gampt. 
rmd.^ 1846y 23^ 38) in order to indicate that the transient initial 
rotatory power of the solution was twice as great as that ultimately 
attained ; this is not so except in the case of glucose, consequently 
the names of '* multirotation " and ' * paucirotation " have been intro- 
duced by Wheeler and Tollens (iinna/«n, 1889, 254, 310) to indicate 
that the ratio of the two rotations is a number greater or less than 
unity. The essential feature of the phenomenon, however, is the 
ehcmge of rotatory power which takes place in the freshly prepared 
solution, and the term ' mviaroi<Uwn * may, therefore, be used with 
advantage to include all those cases in which such a change occurs 
without reference either to the sign or to the relative magnitude of 
the initial and final rotations. 

The phenomenon of mutarotation was first observed by Dubrunfaut 
(JoceU,) in an aqueous solution of glucose, and has since been shown to 
occur in aqueous solutions of eight other sugars, and also in the case 
of several of the acids and lactones derived from them ; a full list of 
^ese substances, together with the limiting values of their specific 
rotatory powers, is given in the accompanying Table I (p. 214). The 
changes of rotatory power which take place on dissolving gluconic 
lactone in water were shown by Fischer to be due to its partial conver- 
sion into the acid in accordance with the reversible equation CqHjo^« + 
Hfi ^nQ^^fi^ ] he, therefore, suggested {Ber., 1890, ^ 2626) 
thaty in the case of the carbohydrates also, the phenomenon might be 
due to hydration^ the aldehyde or ketone being converted into a 
polyhydric alcohol, just as acetaldehyde appears to form an unstable 
hydrate when dissolved in water; - Cflg'CHO + HjO ^ CH8*CH(OB[)2. 
This suggestion of Fischer's is that which has usually been adopted 
to explain the mutarotation of the sugars. The case of ir^bromonitro- 
camphor described in this paper affords, however, a clear case of 
mutarotation which is caused by isomeric change and not by hydra- 
tion, and it therefore appears probable that the mutarotation of the 
sugars is likewise due to isodynamic change, and that when the rotatory 
power has reached its constant value two dynamic isomerides are 
present in equilibrium in the solution. In the case of glucose and of 
lactose, both isodynamic forms appear to have been isolated in an 
wnhydnma eanditian as a- and y-glucose, and a- and y-lactose ; and it 
has been shown that when dissolved in water the two forms ulti- 
mately yield solutions which are similar in every respect, the constant 
value of the specific rotatory power being the same for both, and 
intermediate between the values observed in the freshly prepared 
solutions. The data necessary to ascertain the structure of the 
dynamic isomerid^ are not yet available, but it may be suggested 

Digitized by VjOOQIC 



that, in the case of glucose, they would probably be represented by 
two of the following formuln. 



f (jJHOH 



The explanation of the phenomenon of matarotation given here is 
essentially different from that recently put forward by Trey (Ztit. 
phytikal. Chem., 1895, 18, 193 — 218. Compare lippmann, Ber., 1896, 
29, 203 — 20t), who regards the phenomenon as due merely to a change 
of configuration in the molecule, and not to a change of structure 
such as that suggested above. Moreover, both Lippmann (loe. cit.\ 

Table L—Mutarotatary sugars, 

acids f and lactams. 






+ 106-2** 

+ 62-6 
+ 22*6 

+ 82-9 

+ 66-2 
+ 36 

+ 119 



+ 78-6 

+ 117-6 
+ 167 


+ 62-6° 



+ 65-2 


+ 187 


+ 8-6 

+ 19-2 

+ 80-8 
+ 104-6 


/Parens and Tollens, Ann,, 1890, 
\ 267, 164. 

Tanret, C.R., 1896, 120, 1060. 




a-Lactose hydrate... 

i3-Lacto8e „ 
7-Lactoae „ 


/Pawns and Tollens, Ann., 1890, 
I 267, 170. 

Tschiiioger, Ber,, 13, 1917 ; 14, 
r 2121 ; 26, 1466. 

/ Parens and Tollens, Ann,, 1890, 
I 267, 172. 

Ibid.,^, 166. 
/Schnelle and Tollens, Ann., 
\ 1892, 271, 60. 

/Parens and Tollens, Ann,, 1890, 
I 267, 176. 

Ibid,, p. 168. 

Ibid., p. 174. 
/Gunther and Tollens, Ann., 
\ 1892, 271, 90. 








„ lactone.. 
Glnoonic acid 


+ 66-9 




+ 12 
+ 18-8 


+ 17-6 

/Schnelle and Tollens, Ann,, 
\ 1892, 271, 68. 

Ibid., p. 74. 

Ibid., p. 81. 

/Allen and Tollens, Ann,, 1890, 
\ 260, 212. 

„ lactone 

Galactonic acid 

„ lactone... 


Xvlonic acid 

Digitized by 



and Tanr^t (Btdl. Soe. Chem., 1896, [iii], 15, 195) represent the 
mntarotation as being due to the campl$U change of an unstable into 
a stable modification of the sugar, and therefore describe )3-gluco6e as 
being isomeric with a- and y-glucose; according to the views put 
f<Mrward in this paper, a- and y-glucose and a- and y-lactose represent 
the two iflodynamic forms of the sugar, in a state of greater or less 
pmitj, whilst j9-glucose, )3-lactose, isc,, do not represent any new form 
of isomerism, but are merely mixtures of the isodynamic a- and 
'^forms. The view that the sugars are isodynamic substances in 
exactly the same sense as tr-bromonitrocamphor and ethylic aceto- 
aoetate appears to be entirely new, and the relationship here observed 
between mntarotation and dynamic isomerism may be regarded as a 
general one in all cases in which it can be shown that there is no 
chemical interaction with the solvent such as that which leads to the 
hydrolysis of lactones in aqueous solution. 

MtUarotatian of NUroo(»mphar. 

Ordinary crystalline nitrocamphor melting at 102^ may be regarded 
as oonsisMng of the normal form, its homogeneity being vouched for 
by the constancy of its initial specific rotatory power, and by its well- 
defined, crystalline form. The column headed '' Initial Specific Rotatory 
Power '* in Table II represents, therefore, the effect of the different 
solvents on the specific rotatory power of normal nitrocamphor. The 
most noticeable feature in this column is the sharp contrast which 
exists between the paraffinoid and the benzenoid series of solvents ; 
with the exception of carbon bisulphide, in paraffinoid solvents the 
initial specific rotatory power is in all cases less than 40% the values 
ranging from (P in formic acid to — 37^ in ether ; the benzenoid 
solvents, on the other ^hand, give much larger numbers, the values 
ranging from - 87^ in a 20 per cent, solution in toluene to - 124^ in 
a 5 per cent, solution in benzene. The introduction of a side-chain 
into the benzene nucleus causes a decrease in the rotatory power of 
the solutbn ; thus, in the case of ethylic benzoate, in which the side- 
chain and the nucleus are of nearly equal weight, a value is obtained, 
[aId^* - ^^t which is intermediate between those obtained in the two 
groups of solvents. A further point of contrast is seen in the effect 
of concentration on the initial rotatory power ; in solutions in chloro- 
farm, the values all fall within a degree, but in benzene an increase 
of 15 per cent, in the concentration of the solution causes a decrease 
of 24® in the initial specific rotatory power, and a similar decrease of 
19® occurs in the case of toluene. The differences between adjacent 
members of a homologous series are also much greater in the case of 
benzenoid than of paranoid solvents ; thus, ^ere is a change of 

Digitized by 



+ 18° on passing from benzene to toluene, and a farther difference of 
+ 7° between toluene and xylene, but the difference between ethylio 
and propylic alcohols is only + 2% and acetic and propionic acids differ 
only by «- 2°. These points of difference are not due to the formation 
of a polymeric modification of nitrocamphor in the benzenoid solvents, 
as an increase of concentration would then produce an increase of 
specific rotatory power in the solution instead of the decrease actually 
observed, and, moreover, the freezing point of the solution corre- 
sponds with a normal molecular weight ; they can be best explained 
as being due to the influence of the normal nitro-group on the 
benzene nucleus leading perhaps to the production of an unstable 
additive compound in the solution. This view would be in accordance 
with the known existence of compounds formed by the loose com- 
bination of picric acid with various benzenoid hydrocarbons, and 
would serve to explain the fact that the anomalous behaviour of 
benzene and its homologues is most marked in dilute solutions, that 
is, when the hydrocarbon is present in large excess ; it may also be 
mentioned that jMetMio-ir-bromonitrocamphor has the same rotatory 
power in benzene and in chloroform, whilst the normal a-bromonitro- 
camphor exhibits a difference of 32° in the two solvents. Reference 
may be made here to the difference between the acids and alcohols of 
the fatty series, which is especially marked in the lower homologues ; 
thus formic acid and methylic alcohol give a difference of 31°, which 
falls to 23° in the case of acetic acid and ethylic alcohol, and to 19° 
in propionic acid and propylic alcohol, the two series converging with 
increasing molecular weight towards a value of about - 15°. 

On passing into solution, normal nitrocamphor at once begins to 
change into the pseudo-form, and as the two forms differ very widely 
in rotatory powers this alteration of structure is accompanied by a 
corresponding change in the rotatory power of the solution ; within a 
few days or hours from the time of preparing the solution, the specific 
rotatory power becomes constant after undergoing a change of from 
10° to 24° to the right ; the magnitude of the change in the different 
solvents is given in column 5 of Table II, and is also represented 
graphically in Fig. 1. It has not been found possible to isolate pseudo- 
nitrocamphor in a pure state, and the percentage of each form present 
in the state of equilibrium cannot, therefore, be calculated with any 
degree of accuracy ; as an approximation, its specific rotatory power 
may be assumed to be about + 180°, a figure which would compare 
with that of its anhydride, of which [a]D» +187° in benzene, and 
+ 167° in chloroform, and also with that of its ir-bromo-derivative for 
which [a] » + 189° (see also p. 235), Taking this value as the rota- 
tory power of pseudonitrocamphor, the percentage of this substance in 
the state of equilibrium would be about 7 per cent, in both the 

Digitized by VjOOQIC 



paraffinold and benzenoid solvents. From the values given in chloro- 
form eolations, it will be seen that the percentage of the pseudo-form 
increases with the concentration of the solvent, the change of rotatory 
power being 12^, 14% and 16^ respectively in 5, 10, and 20 per cent. 





-45°^ :^0 

+ 20 

Benzeie 5% 
-» Bi mzene 

Toluenh 5% 

> Berzene 20% 

TolukiJe 10% 




vate 6 

— > Cirbon 

Etier 5% h 
Ethylii^ Acetate 



I alcoHfil 5% 


alcohol 5 % 


Propiohic i 









Fig. 1. — Sbowino magnitudb of Changs of Rotatobt Powbb of 
nitrooamphob in diffbrknt solybntb. 

VdocUy qftlte Isomeric Change. — In most of the solvents examined, 
the whole course of the change of rotatory power has been followed, 
and the observations have -been plotted in the form of curves, such as 
those which are represented in Fig. 2, where the specific rotatory 

Digitized by VjOOQIC 



powers are plotted as ordinates, whilst the absoisssB represent the 
time intervals from the moment when the solvent was added to the 
nitrocamphor. The first reading of the polariscope was usually made 
within three or four minutes, and observations were then taken every 
minute during the first half-hour, and subsequently at longer in- 
tervals, until constant readings were obtained. The earlier part of 
the curve is somewhat irregular as a rule, by reason of temperature 
changes, and the values of the initial specific rotatory power in 
Table II, which were obtained by ezterpolation, are, therefore, given 






tthylK ^e« 


■ t£^) 


— — — " 



\ ^!^ 









Fio. 2. — Cttevbs showing Ohakos op Rotatory Power op Nitrocajiphor 


to the nearest degree, fractional parts of a degree having been omitted. 
The curves given in Fig. 2 are very similar to those which have been 
plotted by Parous and Tollens (Afmalmiy 1890, 257, 160) for aqueous 
solutions of the sugars, and it is shown under ir-bromonitrocamphor 
that the same equation is applicable to both groups of substances. 

Effect qf Concentration, — The isomeric change usually proceeds most 
rapidly in concentrated solutions; thus about a week is required 
before equilibrium is reached in a 5 per cent, solution in chloroform or 
benzene at ordinary temperatures, whilst a few hours are sufiScient in 
a 20 per cent, solution. An opposite result was obtained by Hantzach 

Digitized by VjOOQIC 


and Schnltze {Ber., 1896, 29, 2251) in the case of pseudophenyl- 
nitromethane, but this may be explained by the fact that this com- 
pound is bimolecular, and a dilute solution would, therefore, be most 
favourable to its conversion into the monomolecular normal form. 

^eei qf TemperaUure. — ^The velocity of change is very much in- 
creased by rise of temperature; a 5 per cent, solution in benzene, 
which at ordinary temperatures would change slowly during two to 
four days, was found to have reached a state of equilibrium after 
it had been warmed to 50^ for a quarter of an hour, and then cooled 
during an equal interval. 

If^kionee qf the Sclveni, — ^The interval which elapses before equili- 
brium is reached is shown for a large number of solvents in colomn 6 
of Table II, whilst in column 8 values are given for the period r which 
elapses before the change is half completed. The solvents may be 
divided into two groups. In solvents containing oxygen, such as 
the alcohols, adds, ethers, and ethereal salts, equilibrium is reached 
on the average in about four hours at 15^ the actual periods 
ranging from a quarter of an hour in acetone to half a day in 
ethylio benzoate. The hydrocarbons and compounds such as chloro- 
form and carbon bisulphide are much less active, and a period of from 
two days to a week or more is required before equilibrium is reached 
in these solvents. A similar result was obtained by Hantzsch and 
Schultze {foe. ct^.), who found that peeudophenylnitromethane changed 
more slowly in chloroform and in benzene than in ether, alcohol, 
water, or acetic acid. 

LaiiU SohUianB, — From certain observations that have been made 
with chloroform solutions, it seems doubtful whether the second group 
of solvents are capable by themselves of bringing about the isomeric 
change at all. Out of five 10 per cent, solutions that have been 
examined, three have been found to persist during several days in a 
labile state without undergoing change, and the same result has also 
been noticed once in a 5 per cent, solution, but nothing of the kind 
has been observed in other solvents. Usually, the transference of the 
solution to the polarimeter tube has proved sufficient to start the 
change, but in two teases the labile state was found still to persist • 
however, the change began on warming or shaking the tube. Two 
coirves are given in Fig. 3 (p. 220) to illustrate this ; the first curve re- 
presents the observations made with a 10 per cent, solution, which 
showed an almost constant rotatory power during six days, until the 
labile state was destroyed by warming the tube ; the second curve 
shows the behaviour of a 5 per cent, solution, one portion of which 
was transferred to the observation tube immediately after it had been 
prepared, whilst the second portion was not transfer^red until 16 days 
later ; the initial rotatory power was approximately the same in both 

Digitized by VjOOQIC 


LOWRY: STtJDifiS Of THfi T£ftl»EKES 

cases, indicating that the solution did not undergo any isomeric 
change while in the iBask, but started at once when poured into 
the tube. The dotted line in Fig. 3 (a) shows the normal rate 
of change in a 20 per cent, solution in chloroform. From these 
observations, it appears probable that a solution of pure nitro. 
camphor in pure chloroform might be kept indefinitely without 



o ^r 






L. - 

5 1 



Fio. 8.— Labile Solutionb in Chlorofobm (a) 10 pbr gent., (5) 6 pxb obnt. 

undergoing change; in practice, the solution slowly decomposes 
and assumes a yellow colour, but that even this decomposition does 
not necessarily involve isomeric change follows from the fact that 
a portion of one of the 10 per cent, solutions referred to was kept 
during 63 days in the graduated flask in which it had been prepared, 
and, although it had a marked yellow colour, its rotatory power had 
only fallen from - 27"^ to - 22^ instead of the limiting value - IS"". . 

Digitized by VjOOQIC 


These observations clearly indicate that, at least under certain con- 
ditions, a third substance is necessary to bring about isomeric 
change in the case of nitrocamphor j the results described in the 
following paragraphs serve to suggest that a trace of basic impurity 
is, perhaps, essential as a catalytic agent ; as an alternative hypothesis, 
it might be supposed that a trace of the pseudo-form is necessary to 
start the action ; in either case, there is an interesting analogy to the 
well-known influence of minute traces of water in con^tioning 
chemical change. 

Influence qfBtues, Acids, and Salts. — ^It has been shown by O'Sullivan 
and Tompson (Trans., 1890, 67, 920) that the addition of alkali 
causes the change of rotatory power of the sugars to take place 
instantaneously ; similar effects are produced in the case of nitro- 
camphor, as may be seen from the following experiments. A 
quantity of nitrocamphor was dissolved in chloroform containing 
about 0*5 gram of piperidine per litre ; seven minutes after preparing 
the solution, the speciflc rotatory power was [ajo = - 9^, and no marked 
change of rotatory power was subsequently observed ; the constant 
rotatory power, so rapidly reached in this case, differs from that 
previously given in a 5 per cent, solution in chloroform only to an 
extent which is readily accounted for by the partial conversion of the 
nitrocamphor into piperidine salt. A similar result was obtained 
when nitrocamphor was dissolved in absolute alcohol in which sodium 
had been previously dissolved to the extent of 0*25 gram per litre ; in 
this case, the constant rotatory power had been reached when the first 
reading of the polariscope was taken three minutes after preparing the 
solution. In another experiment, however, in which the quantity of 
sodium was reduced to about 15 milligrams per litre, a very rapid 
change of rotatory power was observed at first, and a constant value 
was reached within 12 minutes from the time when the solution was 
prepared ; the quantity of sodium present in this experiment was too 
small to produce any noticeable effect on the constant rotatory power 
of the solution, which was found to be [ajo" —9^, the normal value 
in a 5 per cent, solution in alcohol. No effect of this kind is produced 
by aeiis \ the addition of 3*5 grams of concentrated hydrochloric acid 
per litre did not cause any increase in the velocity of change of a 
5 per cent, solution of nitrocamphor in alcohol. Neutral salts act 
similarly to free bases, but less strongly ; the addition of 0*7 gram 
of sodium chloride per litre to a 5 per cent, solution of nitrocamphor 
in alcohol was found to reduce the period of change from 5 hours 
to less than an hour. 

The ' catalytic ' influence of bases can be satisfactorily explained by 

equations such as the following, ^^x^^^^^ ^ + NaOEt^ 

Digitized by VjOOQIC 


lowry: stxtdies of the xerpenes 

C8Hh<^6 +H0Et^C3H,^<^i 

+ NaOEt. In these 

equations, the sodium salt is represented as being partially converted 
by the alcohol into sodium ethoxide and nitrocamphor, on the one 
hand, and, on the other, into pseudonitrocamphor and sodium ethoxide ; 
these changes, which, of ^'course, would be incomplete and reversible, 
might be expected to take place with much greater velocity than the 
simple isomeric change, and it may even be questioned whether the 
latter is capable of taking place at all apart from such ' catalytic 
agents.' The interactions which are here assumed to occur in an 
alcoholic solution of nitrocamphor are strictly analogous to the hydro- 
lysis of a salt into its acid and base, which is most marked in dilute 
aqueous solution, and when either acid or base is feeble, for example, 
KCN + HjO^KOH + HON. Similar equations may be used to explain 
the action of the nitrogen bases, but here the change would be analogous 
to the dissociation of ammonium chloride in the gaseous state, or of 
aniline acetate in solution in benzene. The influence of neutral salts 
may be explained as dependent on the * ratio of distribution ' of the 
base between the nitrocamphor and the acid of the neutral salt. It is 
important to notice that these catalytic agents do not disturb the 

Table IL — MutaroUriton of nitrocamphor. 

' Specific rotatory power, 










k, + k^ 















4 days 


: -116 

! -100 

! -106 







4 days 





4 days 



26 h. 

1-8 h. 

4-7 h. 

4-5 h. 

41 h. 
23 h. 

4 1 h. 









Ethvlic benzoate.. 


Carbon bisulphide 





2 days 


6-9 h. 









1 week 
6 hours 

2 hours 



32 h. 
48 h. 

1-9 h. 





Ether , 


Ethylic acetate ... 





2 honrs 


0-2 h. 

4 1 




+ 8 


i hour 


0-06 h. 


Methylic alcohol.. 





3 hours 


0-6 h. 


Ethylic „ 





5 hours 


1-1 h. 


Propylic „ 





2 hours 


0-6 h. 


Formic acid 



+ 12 


3 hours 


0-7 h. 


Acetic ,, 



+ 8 


5 hours 


10 h. 





+ 5 


2 hours 


0-3 h. 


Digitized by VjOOQIC 


equilibriom between the isodynamic fonxui of nitrooamphor, as it has 
recently been asstuned by Sohiff that agents of this kind are capable 
of effecting the complete conversion of ethylic acetoaoetate into one or 
other of its isodynamic forms ; this point will be referred to again 

MiUaraUUion qf wBromonitroeamphor. 

Through the kindness of Dr. Slipping in providing me with the 
necessary material, I have been enabled to extend the investigation of 
nitrocamphor by studying its tr-bromo-derivativei The introduction of 
the bromine atom does not produce any great difference in chemical 
properties, but completely alters the physical properties of the sub- 
stance, the most important result being that the pseudo-form is 
rendered much less soluble and can readily be isolated, whilst it is still 
possible to isolate the normal form. The condition of equilibrium can, 
therefore, be approached from either side, and as the rotatory power of 
the two forms can be determined, the exact percentage of each when 
equilibrium is reached can be ascertained. 

v-Bromonitrocamphor was prepared and examined by Lapworth and 
Sapping (Trans., 1896, 09, 304), and was described by them as being 
trimorphous, the three modifications melting at 108°, 126°, and 142°. 
Two of these forms were submitted to crystallographic measurement, and 
the axial ratios and the drawings of the crystals are given in the paper 
referred to. From the results obtained in studying nitrocamphor, it 
appeared to be possible that the three crystalline modifications 
might include the normal and pseudo-forms of the substance, and they 
were, therefore, submitted to polarimetric investigation in order to 
ascertain the nature and magnitude of any changes which might take 
place after dissolving them. 

The orthorhombic form, melting at 142°, is readily obtained in good 
crystals from solution in chloroform ; when it is dissolved in benzene, the 
value of [a]i> changes from + 188° to - 38° in a 3*33 per cent, solution at 
13° ; in a 5 per cent, solution in chloroform at 14° the limits were + 189° 
to + 31° ; the magnitude of the change is, therefore, very large, being 
two or three times as great as that recorded in any previous 
case. The high dextrorotatory power of this modification indicates 
that it is the ptetulo-iormf whilst its purity is vouched for by its definite 
crystalline form, since the isodynamic forms are not isomorphous. 
That the salts are derived from this modification of the substance is 
shown by their similar dextrorotatory power, that of the hydrated 
potassium salt, CioHjgBrNOjK + 2H2O, being [aj^ - + 221° in a 3 per 
cent, aqueous solution at 15°. 

The tetragonal form, melting at 108°, is difficult to isolate in any 
quantity, as it does not begin to separate until after the ortho- 

Digitized by VjOOQIC 


rhombic form, when the solution in chloroform is allowed to evaporate, 
but the crystals can be readily distinguished and picked out mechanic- 
ally. When dissolved in benzene, this modification shows a change of 
rotatory power from - 61^ to - 38° in a 3*33 per cent, solution at 15° ; 
it, therefore, resembles nitrocamphor both in the magnitude and in 
the direction of the change and must be regarded as the normal form 
of the substance. 

The third modification, melting at 126°, separates from a hot solution 
in spirit as a crystalline powder ; when dissolved in beozene, a specimen 
of this form showed a specific rotatory power changing from + 150° to 
— 39°, and it must, therefore, be regarded as a mere mixture of the 
normal and pseudo-forms similar to that which is obtained on allowing 
a solution in chloroform to evaporate to dryness. Similar mixtures 
are almost always obtained when crystallisation takes place rapidly, 
and it is only by obtaining the substance in well-defined crystals 
that the purity of the two forms can be ensured. 

When brought into contact with a solvent, the different modifica- 
tions of ir-bromonitrocamphor yield solutions which are ultimately 
identical ; in each case, the same constant rotatory power, [ajp » - 38°, 
is reached ; this affords a proof that the equilibrium is the same in 
each case and corresponds to a mixture of 1 part of the pseudo-form 
with 17 parts of the normal. The production of such an equilibrium 
has already been indicated in connection with the different modifica- 
tions of glucose and of lactose, and also, somewhat imperfectly, in 
connection with the acids and lactones referred to in Table I. The 
changes of rotatory power of normal and pseudo ?r-bromonitrocamphor 
in solution in benzene are represented graphically in Fig. 4. 

Normal and pseudo-ir-bromonitrocamphor differ widely in physical 
properties ; they crystallise in different systems, their melting points 
differ by 34°, and the specific rotatory powers by 239° ; the difference 
in solubility is also very marked, for whilst the normal form dissolves 
instantly in cold chloroform, the pseudo-form only dissolves slowly on 
boiling. In general, the normal form shows a behaviour which 
indicates a less solid cohesion ; thus it has a lower melting point and 
a greater solubility and is much softer than the pseudo-form. 

Form qf the Muta/roiation Curves, — ^The change of A^bromonitrocam- 
phor from the normal to the pseudo-form, or viee verad, is a mono- 
molecular action, and might, therefore, be expected to proceed 
according to an equation of the type 

dx/dt = ki{xQ - Xt) - k^ 

^l^ich, when integrated, becomes 

*,+*,-! log 5^, 

Digitized by VjOOQIC 



where k^, k^ are the velocity-constants of the isomeric change in the 
two directidns, and x^, xt, and «qo are the quantities of one of the 
forms at times 0, t, and oo . This equation has already been applied 
by Kiister to the isomeric change of the two hexachloroketocyclo- 


+ 160- 



T lUW 










3. DAYS 

Fio. 4. — Changes or Rotatort Power of Psbudo- and Normal 
v-BROMomTRooAMPHOR IN Bbkzenb (8*88 per cent.). 

pentenes {Zeit. phynkal. Chem,, 1895, 18, 171), and by Trey to the 
mutarotation of the sugars {ibid,, p. 198). In the case of nitro- 
camphor and ir-bromonitrocamphor, the most convenient unit of 
measurement is that given by the specific rotatory power of the 

„.gitized by Google 



solution^ and if oq be used to represent the rotator j power of the 
solution when first prepared, at that at time t, and a^^ the constant 
value of the rotatory power, then the equation maj be written 

-(^ + *2) = 7 {log (ao-aoo)-log (a«-aoo) j , 

where c represents the base of the natural system of logarithms, and 
has the valtie 2 '7183, whilst k^ and A^ are the velocity-constants of 
the isomeric changes in the two directions, and represent the pro- 
portion of each that undergoes isomeric change in the course of one 

hour. In Table III, the values of the constant ^ ^ in the case of 

Tablb m. — Showing change of rotatory power of peeudo^-hroTnonUro- 
cam/pkoT in a 5 per cent, solution in chloroform at 14^. 

at - tt^. 

t (hours). 


*! + *« 







0-2 h. 

+ 188-5' 





0-6 h. 






10 h. 






8-0 h. 






6-0 h. 






7-0 h. 






24 h. 






72 K 






81 h. 






96 h. 






Limit 168 h. 




Mean value 

a 5 per cent, solution of pseudo-Tr-bromonitrocamphor in chloroform 
are given in the sixth column, and their constancy shows that, after 
the first hour, the change proceeds quite regularly according to the 
monomolecular law; the values given in the fourth column for 
at - a^Q are calculated from the equation 

log (at - aoo ) = 0-1981 - 1 x 0-0197, 

and show a very satisfactory agreement with the observed rotations, 
the difEerences in each case being less than 2°. In Table lY, values 
are given for both the normal and pseudo-forms of «-bromonitrocamphor 

Digitized by 




Table IV. — Showvng ckomgt qf roUxUyry power of pemtdo- cmd normal 
ir-bromonitroccMnphor in a 3 '33 per cent. soltUion in benzene. 

t (hoara). 


a« - a«. 





( + 188-4'0 

(220 ^^ 



0-6 h. 

-1- 184-0 




1-0 h. 

+ 1797 




1-6 h. 

+ 176-0 




8-5 h. 

+ 168-2 




6-0 h. 

+ 146-2 




7-0 h. 

+ 129-6 




8-7 li. 

+ 116-0 




27-0 h. 





29 h. 

+ 18-7 




81 -711. 

+ 10-7 




48*6 h. 






Limit 99-3 h. 







Mean value 

Mean valae 










25 h. 




48 h. 




72 h. 




118 h. 





Limit 144 1l 




Mean valae 

in a 3*33 per cent, solution in benzene ; in the case of the pseudo-fomii 
the Telodty-oonstant ^ ^ in column 4 shows a steady increase as 

the action proceeds, but this is due almost entirely to the slowness 
of the isomeric change during the first hour or two, and the later part 
of the curve gives a satisfactory agreement, as is seen from the 
constancy of the numbers given in the fifth column, which are 

calculated from the formula ^L±^ = - / 03700 - log (ac - aoo ) | . Un- 
der comparable conditions, the velocity-constant ^ ^ should be the 

same for both modifications of a substance, but it is only rarely that 
this agreement can be realised experimentally ; in Table lY , the mean 

Digitized by VjOOQ IC 


value 0*0188, in the case of the pseudo-form, is three times as great as the 
valuoi 0*0064, obtained in the case of the normal form ; this difference 
maj be accounted for in part bj the difference of 2° in temperature in 
the two experiments, but it is probable that some other points of 
divergence existed between the conditions in the two experiments. 

If 0*0188 be taken as a mean value for Vt^ then Jfci+ifej = 00511, 

and since Ajj rAij- 226*4: 13*9, *i = 0*0481 and Aig^ 0*0030, where k^ is 
the velocity-constant of the change from pseudo- to normal, and k^ 
that of the change from normal to pseudo- ; thus, under the conditions 
of the experiment, 4*8 per cent, of the quantity of the pseudo-form 
present in the solution would undergo isomeric change in one hour, 
and only 0*30 per cent, of the normal form. 

In the case of nitrocamphor, the changes of rotatory power are too 
small to give a good result on calculating the velocity-constant from 
the separate observations, but in Table II the period r which elapses 
before the change is half-completed is given in column 8, and from 
this the constant k^ + k^ has been calculated for each solvent from the 

formula kj^ + k^-- log |. 

Campariaon with the NUropairaffina. 

It was, I believe, first suggested by Dr. Armstrong in the year 
1892 (Proc., 1892, 101), that the salts of the nitroparaffins are derived 
from 'a hydroxylic modification of the substance, and not directly 
from the nitro-compound. This suggestion was verified shortly after- 
wards by the experimental work of Nef {Annalen^ 1894, 280, 263 ; 
Ber.f 1896, 29, 1218), and since then a number of attempts have 
been made to isolate the hydroxylic form of various nitroparaffins. 
Labile crystalline modificationsof phenylnitromethane, C^H^' CHg* NOj, 
and parabromophenylnitromethane, C^H^Br'CHj'NOj, have been 
prepared by Hantzsch and Schultze {Ber., 1896,29, 699 and 2256) by 
acidifying an aqueous solution of the sodium salt with mineral acids, 
and the same method has been successfully used by Konowaloff in the 
case of iD-nitromesitylene, C^HgMeg'OHj'NOj, diphenylnitromethane, 

g«^»>CH-NOj, isopropylphenylnitromethane, q«2''>CH-N0„ and 

other secondary nitroparaffins {Ber.^ 1896, 29, 2193). These labile 
forms appear to consist largely of the dynamic isomeride from which 
the salts are derived, but ir-bromonitrocamphor seems to be the 
only nitro-compound of which the two isodynamic forms have been 
obtained in well-defined crystals of ascertained purity, and submitted 
to crystallographic measurement (L. and El., loe. oil.). The results 

Digitized by VjOOQIC 


obtained in the two series of nitro-com pounds differ in several 
important respects, and the experiments described in the following 
paragraphs are, therefore, given in the form of a comparison. 

Action qf Adda on Aqueous SoluUona qf the Salts. — ^The labile 
peeudo-forms isolated by Hantzsch and Schultze were prepared by 
the action of mineral acids on a solution of the sodium salt, whilst 
organic acids, such as carbonic and acetic, on account of their slower 
action, were found to give either the normal form or a mixture of 
the two. As this method appeared to afford a ready means of pre- 
paring pseudonitrocamphor and normal n^bromonitrocamphor, attempts 
were made to isolate these. On acidifying a dilute solution of potas- 
sium n^bromopseudonitrocamphor, a crystalline powder was obtained 
which showed an initial specific rotatory power [a] =+116° in a 
3-33 per cent, solution in benzene; the product was, therefore, a 
mixture containing only 30 per cent, of the pseudo-form. Moreover, 
on attempting to isolate pseudonitrocamphor by the action of mineral 
acids on a solution of one of its salts,'a mixture was again obtained in 
every experiment, the proportions of the two forms being approxi- 
mately the same as that in which they exist in solution; in one 
experiment, a known weight of nitrocamphor was dissolved in aqueous 
potash, precipitated with hydrochloric acid, drained on a filter-pump, 
and immediately dissolved in a known volume of alcohol ; the solution 
gave a constant rotatory power, [a]^ » - 9° ; the first reading of the 
polariscope was obtained within 15 minutes from the time when 
the acid was added to the salt solution, and no change was subse- 
quently observed. If the pseudo-form be assumed to be the first 
product of the action, the production of a mixture of the two forms 
can be readily explained as due to the catalytic influence of the salts 
present during precipitation (compare p. 221), but it is also possible 
that both forms are produced directly from the potassium salt by the 
action of the acid, without being first thrown down in the pseudo-form. 

Nature qf the Isomeric Chofnge^ Complete or Incomplete, — It has been 
stated by Hantzsch and Schultze that the conversion of the pseudo- 
into the normal nitroparaffins is complete, that is, that the isomeric 
change is non-reversible; '* whilst thus the groups -CHj* CO- and 
-CHIC(OH)- are both stable to a certain extent in the free state, of 
the groups -CHg-NOg- and -CHINOsH, the first alone is stable" 
(ffttti., p. 2267). This statement is directly opposed to the equilibrium 
which has been shown to exist between the two forms of nitrocamphor 
and ir-bromonitrocamphor, and is also contradicted by the statement 
of Konowaloff {loe. ct^.), that the labile pseudo-form of xylylnitro- 
methane, which only differs from phenylnitromethane by the presence 
of two additional methyl groups in the benzene nucleus, is converted 
into a mixture of the normal and pseudo-forms, containing also a 

Digitized by VjOOQIC 


small amount of the corresponding aldehyde. It appears, therefore, 
to be more probable that, even in the case of phenylnitromethane and 
parabromophenylnitromethane, an equilibriom is established between 
the two forms in solution and in the liquid state, although the 
quantity of the pseudo-form may be too small to produce any marked 
coloration with ferric chloride. 

Action qf Ferric Chloride. — ^In almost every case of dynamic iso- 
merism that has been studied up to the present, the formation of a 
coloured ferric salt has been made use of to distinguish the hydroxylic 
modification, and it has even been made the basis of rough quantitative 
measurements. In the case of nitrocamphor and x-bromonitro- 
camphor, this test is not available, as both isodynamic forms give the 
same coloration with ferric chloride ; in order to make sure of thus 
result, crystals of normal and pseudo-7r-bromonitrooamphor were 
crushed and dissolved in equal quantities of cold spirit, and a drop 
of alcoholic ferric chloride was immediately added to each ; not the 
slightest difference could be detected between the two solutions, 
either in the depth of colour or in the velocity with which it was 
produced. This unusual behaviour is most readily explained as being 
due to the accelerating effect of ferric chloride on the isomeric change, 
an effect which has already been shown to be produced by neutral 
salts ; it would, therefore, appear that, under the influence of the ferric 
chloride, equilibrium is reached during the second or two that is 
required for the development of the coloration. 

Action of Nitrous Acid on NUroeampkor. — It has been stated by 
Hantzsch and Schultze that, when suspended in ether, all pseudo- 
nitro-compounds and their derivatives yield a' sky-blue coloration if 
acted on with hydrogen chloride or acetic chloride. This does not 
occur in the case of nitrocamphor, but a similar coloration can be 
produced by the action of nitrous acid. If a drop of dilute sulphuric 
acid is added to an aqueous solution of potassium pseudonitrocamphor 
and potassium nitrite, a transient blue colour is produced, which 
rapidly passes to green and then to yellow ; the ultimate product is a 
yellow oil, which decomposes almost explosively when heated, giving 
off a considerable quantity of gas and leaving a solid cake of oamphor- 
quinone. From analogy to other secondary nitro-compounds, it is 
probable that the blue substance is the peeudomtrcl^ and that this 
passes by loss of 2N0 into camphorquinone, as shown by the formulas 

The blue coloration observed by Hantzsch and Schultze must obvi- 
ously be due to the nitrous acid which is liberated by the decomposition 
of the unstable pseudonitroparaffins.^ 

Digitized by VjOOQIC 


Mokevla/r Weights. — It has been shown by Hantzsch and Schnltze 
that pseudophenylnitromethane is largely bimolecnlar, and that when 
it is dissolved in benzene the molecular weight, as determined from 
the freezing point of the solution, falls slowly from 210 to 141 during 
the period that elapses before equilibrium is reached ; the molecular 
weight of the normal form, calculated from the formula C^HYNOg, is 
137. A similar decrease of molecular weight has been shown to occur in 
a freshly diluted solution of formaldehyde (Eacheweiler and Groesmann, 
Annalen, 1890, 258^ 103). On making an experiment with pseudo- 
ff^bromonitrocamphor, the freezing point of the freshly prepared solu- 
tion was found to be 0*492^ lower than that of the benzene used as 
a solvent; when the solution had been brought to a condition of 
equilibrium by keeping it at atmospheric temperature during 6 days and 
then momentarily warming to the boiling point, the depression of the 
freezing point was 0'652% the difEerence between the two values being 
no greater than that which might be produced by the evaporation 
of the benzene. The conoentration of the solution was found to be 
2'946 grams per 100 grams of benzene, and the calculated molecular 
weight of the substance was 261 ; from this, it may be concluded 
that both the isodynamic forms of ir-bromonitrocamphor are mono- 
molecular, the theoretical value for a single molecular proportion 
being M» 271. 

A determination of the molecular weight of nitrocamphor in solu- 
tion in benzene, by the freezing point method, gave the value M => 208 
(calo. 107). 

The CryetoMieation qf Dyiwmic leomerides. 

CryftaUieation frcm SohUion. — ^The behaviour of ir-bromonitro- 
camphor affords a good illustration of the laws which govern the 
crystallisation of isodynamic substances, and as these do not appear 
to have been previously discussed, a short account of them is given 
here. For each form there will be a true eolubilUy which is inde- 
pendent of the isomeric change, but as the isomeride is also present 
in the solution when equilibrium is reached, the quantity of the sub- 
stance ultimately dissolved, or the apparefnA aolvbUiiyy will always be 
greater. Thus, every gram of pseudo-n^bromonitrocamphor is in 
equilibrium in solution with 17 grams of the normal form, and the 
apparent solubility will, therefore, be 18 times as great as the true 
solubility, whilst for the normal form the ratio will be 18/17. In the 
ease of w^bromonitrocamphor, the apparent solubilities of the two forms 
are very nearly the same, as a mixture of the two is usually obtained 
on crystallisation ; that of the normal form, however, is slightly the 
larger, as the pseudo-form is always the first to separate from solution ; 
for the sake of definiteness, it may be assumed that the true solu- 

Digitized by VjOOQIC 


bilities of the normal and pseudo-forms are in the ratio of, say, 19 : 1, 
in which case the apparent solubilities would be in the ratio of 19 : 17. 
On allowing a solution of ii^bromonitrocamphor in chloroform to evapo- 
rate, the pseudo-form is observed to separate first, and as the isomeric 
change always proceeds in such a direction as to maintain the equi- 
librium, the whole of the substance can be obtained in this form if 
the crystallisation is sufficiently slow. If^ however, evaporation takes 
place more rapidly, the quantity of the pseudo-form in the solution 
will decrease, and, when the proportion falls to lees than one-twentieth 
of the total quantity of substance in solution, the normal form will 
begin to separate, and both forms will then crystallise out side by 
side ; for this reason, it has been observed that the normal form only 
separates in the last stages of the evaporation, and it is only the com- 
plete disappearance of the solvent that saves it from being re-dissolved 
and converted into the less soluble pseudo-form. 

In the case of a substance like nitrocamphor, in which the form that 
predominates in solution is the first to separate, there will not be any 
marked decrease in the proportion present in the solution unless the 
crystallisation is extremely rapid, and even then the pseudo-form can 
only separate in a very small proportion, and would appear merely as 
an impurity in the normal form. For this reason, it is very easy to 
obtain normal nitrocamphor in a pure state, and no indication is 
obtained of the presence of the pseudo-form in the crystals even when 
the substance separates from solution very rapidly. 

Fusion. — When ?r-bromonitrocamphor is carefully fused on a glass 
slide and allowed to cool, crystallisation sets in, and when examined 
under a microscope between crossed Nicols, the slide is seen to be 
covered with a brilliant, microcrystalline figure ; the appearance of 
the image varies considerably, but radiating lines of minute vacuum 
bubbles, caused by the contraction of the substance on solidification, 
are usually present, and appear to be somewhat characteristic of the 
substance. When examined with an objective of higher power, bi- 
axial interference figures are seen, which show a + double refraction ; 
the substance that crystallises out under these conditions is, there- 
fore, the pseudo-form, since the normal form gives a uniaxial inter- 
ference figure with, double refraction (L. and K., loc. cit.) ; the 
separation of the pseudo-form from a mixture which probably consists 
chiefly of the normal form, is a somewhat striking illustration of the 
difference* in their^ mutual solubility, and recalls the similar result 
obtained on crystallising from solution. Under ordinary conditions, 
a mixture of the two forms is obtained on cooling the fused sub- 
stance, just as a mixture is obtained on rapid crystallisation from 
solution ; a specimen that had been fused in a test-tube and rapidly 
cooled showed an initial specific rotatory power [a]|>= -i-3^ in a 3*30^ 

Digitized by VjOOQIC 


per cent, solution in benzene at 17% and therefore contained 23 per 
cent, of the pseudo-, and 77 per cent, of the normal form ; this result 
does not indicate the composition of the fused substance^ but since 
the pseudo-form is the first to crystallise, it may be regarded as 
giving a maximum limit to the percentage of this in the liquid 

After fusion, all the modifications of n^bromonitrocamphor remelt at 
126^ (L. and K., he. eit.\ and this temperature, therefore, must repre- 
sent the point at which the solid pseudo-form can exist in stable 
equilibrium with the liquid mixture. At this temperature, the 
solubility of the pseudo-form in the fused normal form must be 
exactly equal to the percentage in which it exists in equilibrium with 
it in the liquid state ; in the case of ir-bromonitrocamphor, this point, 
which maybe called the '* equilibrium temperatuo'e,^' lies 16° below the true 
melting point of the pseudo-form. The true melting point of normal 
v-bromonitrocamphor, which is given by Lapworth and Kipping as 
108°, lies below the equilibrium temperature; but it is somewhat 
difficult to observe this melting point, as the fused substance rapidly 
undergoes isomeric change and is converted into the solid pseudo-form, 
after which it melts again at 126°. 

When fused nitrocamphor is allowed to cool, it does not crystallise, 
but solidifies in an amorphous condition, indicating that the mutual 
solubility of the two forms is greater than in the preceding case. The 
initial specific rotatory power of a recently fused specimen was found 
to be [a]D= -7° in a 5 per cent, solution in chloroform at 17°, 
corresponding with the presence of about 10 per cent, of the pseudo- 
form ; as the forces of crystallisation do not, come into play, it may 
be assumed that these numbers correspond also with the composition 
of the fused substance. The condition of equilibrium in solution is 
reachedy on the average, when about 7 per cent, of the substance is in 
the pseudo-form, and this experiment shows that, at a temperature 
some 80° higher, the proportion of the pseudo-form present in the 
fused product is not materially greater, since the figure given above 
is probably too high on account of the formation of the anhydride 
during fusion. The presence of both normal and pseudo-nitrocamphor 
in the liquid substance, and their great mutual solubility, affords a 
ready explanation of the large amount of oil that is obtained in 
preparing it, for a relatively small amount of a third substance, 
present as an impurity, would suffice to prevent the mixture from 

LimiUa of StdbilUy. — ^The term '' stability limits," or *' cippareTU melting 
pakU," has been introduced by Knorr {Annahny 1896, 293, 70 and 88) to 
indicate the temperature at which the solid substance undergoes 
isomeric change, the product being usually a liquid mixture of the two 

Digitized by 



forms. In the case of normal ?r-bromonitrocamphory the substance 
only melts when heated rapidlj to a temperature of 108^ ; when heated 
more slowly, it is converted into the solid pseudo-form, and then melts 
at 126^; for this substance, the stability limit is probably just below 
the true melting point. In the case of normal nitrocamphor, the 
observed melting point is by no means sharp or well-defined, and may 
be regarded as merely a stability limit. For pseudo-ir-bromo- 
nitrocamphor, the stability limit is identical with the ' equililnrium 
temperature ' at which the substance remelts after fusion, but this 
will not necessarily be so in other cases ; thus there would be a definite 
stability limit or apparent melting point for pseudo- as well as for 
normal nitrocamphor, but there is no equilibrium temperature for 
this substance, as it does not crystallise again after fusion ; thus the 
equilibrium temperature always represents a reversible change, whilst 
the stability limit represents a non-reversible change in the case of one 
of the isodynamic forms, and often in the case of both. The in- 
stability of the solid pseudo-forms of phenylnitromethane and para- 
bromophenylnitromethane can readily be explained by means of the 
stability limits and the equilibrium temperature ; as isomeric change 
takes place at the ordinary temperature, the stability limits of both 
pseudo-forms must obviously lie below the atmospheric temperature, 
but whilst parabromophenylnitromethane has an equilibrium tempera- 
ture at 60°, and therefore passes without melting into the solid normal 
form, that of phenylnitromethane is below the atmospheric tempera- 
ture, and a liquid is produced as the change takes place. 

Vaport8atian» — So far as I have been able to ascertain, the isodynamic 
forms described in this paper are quite stable in the crystalline 
condition ; it must, however, be remembered that a very small 
quantity of solvent would be sufficient to convert a large amount of 
the more into the less soluble form, and the complete conversion of 
(e.g.) normal into pseudo-ir-bromonitrocamphor might even be effected 
by the vapour of a solvent; moreover, the conversion of the labile 
into the stable form must, in any case, take place more or less 
rapidly by a process of sublimation. Just as in the case of solubility, 
there will be a true and an appcMr&nt vapour pressure, and the labile 
form will be that which has the greater apparent vapour pressure ; in 
the case of ?r-bromonitrocamphor, this is evidently the normal form 
which has the greater apparent solubility and the lower melting 
point ; the following experiment appears to show that this is also the 
case with nitrocamphor, and that the conversion of the normal into 
the pseudo-form actually takes place by a very slow process of 

It was noticed that a highly purified specimen of nitrocamphor, when 
kept in a stoppered bottle for several months, acquired a sweetish 

Digitized by VjOOQIC 


Bmelly whilst the sides of the bottle became clouded over with a thin, 
white snblimate ; the amount of substance thus formed was exceed- 
ingly small, but when the bottle was rinsed out with chloroform, a 
dextrorotatory solution was obtained, which showed a change of 
rotatory power from right to left ; this behaviour is what might be 
expected f roih pseudonitrocamphor, and a rough estimate of the initial 
specifie rotatory power gave the value [a]o = + 160° ( ± 10°), which is 
slightly less than the values previously assumed for this substance and 
observed in the case of its anhydride; that the sublimate does not 
consist of this anhydride is shown by the fact that it does not become 
yellow on exposure to light, and also by the changing rotatory power 
of the solution. The evidence brought forward here is admittedly 
slender, but the observation seems to demand some such explanation 
as that given above. 

Dynamic Isomerism. 

The isomerism of normal and pseudo-^ir-bromonitrocamphor and of 
normal and pseudonitrocamphor is found by experiment to be of such 
a nature that the two forms can only exist separately in a stable con- 
dition when in the solid state, and then only at temperatures below 
the stability limit ; on passing into the fluid state, either by fusion or 
by dissolution or by vaporisation, a reversible isomeric change sets in, 
and after a longer or shorter period of time a state of equilibrium is 
reached in which equal numbers of molecules of the normal and 
pseudo-form undergo isomeric change in any given interval of time. 
For this well-defined type of isomerism, it appears advisable that the 
term ** dynamic isomerism'* should be used as being more significant 
than the other terms which have been proposed, and in order to recall the 
dynamic equUihritMn which exists between the isomerides in the fluid 
state ; two substances which behave in this way may be described as 
being ' isodynamic,' or they may be referred ^to as 'isodynamic 
forms ' of one substance (see the article on " Isomerism " by Dr. 
Armstrong in Morley andMuir's edition of Waifs Dictumary qf 
Chemistry), In the case of ^bromonitrocamphor, equilibrium is 
reached in solution when 6 per cent, of the substance is in the pseudo- 
form, and the isomeric change from the pseudo- to the normal form 
must, therefore, proceed 16 times as fast as the change in the reverse 
direction ; from the velocity-constants given on p. 228, it is seen that in 
eaehhour 4'8 per cent^ of the pseudo-form and 0'30 per cent, of the 
normal form underwent isomeric change under the conditions of the 
experiment. In the case of the separate molecules, the structure will 
at one tame be that of the normal form, and at another that of the 
pseudo-form, but the isomeric change will only take place somewhat 

Digitized by VjOOQIC 

236 lowry: studies of the terpenes 

slowly ; thu8| in the experiment referred to, the average duration of 
the two phases would be 21 hours and 13 days respectively.- 

A theory similar in many respects to that of dynamic isomerism 
was put forward in the year 1885 by Laar in a paper, " Ueber die 
Moglichheit mehrerer Strukturformeln fur dieselbe chemische 
Yerbindung" {Ber., 18, 648—657). In order to explain the fact 
that many chemical compounds are capable of yielding derivatives of 
two types, and that conversely many interactions which might be 
expected to yield isomerides actually give identical products, he 
assumed that, in all such cases, the * tautomeric ' substance alternated 
backwards and forwards between the configurations represented 
by the two alternative structural formulsB, so that at one 
moment it had the structure represented by one formula, and at 
another moment that represented by the other. Thus it has been 
found that the same substance is produced by the action of nitrous 
acid on phenol and of hydroxylamine on quinone, and Laar, therefore, 
assumed that the compound was continually alternating between 

the phases \ f and I | . The essential difference between 

^"0 i^OH 

Laar's theory of tautomerism and that of dynamic isomerism con- 
sists in the fact that he regards the phenomenon as intramolecular, and 
not intermolecular, comparing it with the vibration which gives rise to 
the spectra of incandescent gases and contrasting it with the equi- 
librium of dissociation. Two consequences follow from this : first, he 
regards all isodynamic bodies as homogeneous substances having uniform 
properties, whereas, in many cases, they can be proved to be mixtures 
of two isomerides which have different physical properties; second^ 
the phases of the substance could only be separated by cooling to the 
absolute zero, and so stopping all motion in the molecule, whereas 
actually it is found to be sufficient in most cases merely to bring the 
substance into a crystalline form. For this reason, Laar's speculative 
theory has been generally abandoned, whilst the theory of ' dynamic 
isomerism ' has been verified in a large number of examples. 

The term * desmotropy* was introduced by Jacobson (Ber.f 
1888, 21, 2628) as an alternative to the word * tautomerism,' but its 
meaning was narrowed down by Hantzsch and Hermann {Ber.^ 1887, 
20, 2802), who suggested that the term should be used only for 
those substances of which two solid modifications are known. Two 
consequences have followed from this : first, a considerable amount of 
confusion has been introduced in relation to the difference betweei) 

Digitized by VjOOQIC 



polymorphism and dynamic isomerism, and second, according to the 
present style of nomenclature, ^^bromonitrocamphor would be 
described as ^ desmotropio/ whilst nitrocamphor would be merely 
* tautomeric,' although the isomerism of the normal and pseudo- 
forms is of precisely the same nature in each case. 

Two years previous to the publication of Laar's paper, it was 
shown by Baeyer (^er., 1883, 16, 2188) that certain compounds of 
the indigo-group were capable of yielding two types of derivatives, 
although they were not themselves known in more than one of the 
two possible isomeric forms. He, therefore, assumed that one of the 
isomerides was stable, whilst the other was labile, and went over 
immediately into the stable form whenever attempts were made to 
prepare it. Thus, pseudoisatin is not known, although ethyl pseudo- 
isatin is perfectly stable, and, similarly, pseudoindoxyl is known only 
in the form of derivatives. 

C,H,-CO <?,H,-CO C,H,- 

N=rr:COH NH CO NEt- 

Isatin (stable). Fseadoisatin (labile). Stable derivatiTe. 

C,H,-j:]OH VeH^-gO (f,^,-QO 

Indozyl (stable). Pseadolndozyl (labile). Stable deriyatiye. 

Baeyer's hypothesis, to which the name of ^ paeudomeriam* has been 
given by Laar {Ber,, 1886, 19, 730), may be regarded as contemplating 
merely a limiting case of ' dynamic isomerism ' in which the pro- 
portion of one of the isodynamic forms in the state of equilbrium is 
infinitesimal, so that the isomeric change is practically complete, and 
proceeds only in one direction. 

The credit of recognising the possibility of dynamic isomerism and 
of applying it to explain the behaviour of substances, such as hydro- 
cyanic acid, which yield two kinds of derivatives, undoubtedly 
belongs entirely to Butlerow. In a paper, ''Ueber Isodibutylen •" 
(Annalen, 1877,189, 44), he showed that the products of oiddation of the 
substance indicated the presence in it of two isomeric hydrocarbons, 

^>CH- CH^- CH:C<Jg3 and ch'>^^' ^^a' OH^- (K^J^, 
and brought forward evidence to show that, in presence of concen- 
trated sulphuric acid, there is a state of equilibrium between the two 
olefines, the intermediate stage being probably the alcohol or alcohol 

sulphate, ^^3>OH-CH2-CH-C(SO^H)<^g8. He, therefore, sug- 


gested that, in other cases, a similar state of equilibrium might exist 
even in the absence of a catalytic agent, and pointed out that this 
would fully account for the production of two series of ethers from 

Digitized by VjOOQIC 

238 lowry: studies of the tebpenes 

hydrocyanic and other allied acids; the recent statement of Knorr 
{Ber.y 1897, 30, 2389), that hydrocyanic acid and other ' tautomeric ' 
liquids are composed of a mixture of two isomerides in equilibrium, is, 
therefore, identical in every respect with the idea so clearly stated by 
Butlerow 22 years previously. 

To Butlerow belongs also the credit of recognising the importance 
of catalytic agents in effecting isodynamic change. These substances 
at least increase its velocity, if they do not underlie the isomeric change, 
a result which is seen very clearly in the action of bases in accelerating 
the changes of rotatory power of nitrocamphor and the sugars ; in 
many instances, therefore, they render obvious cases of isodynamic 
change which would otherwise proceed too slowly to be observed. To 
this class of substances, Butlerow's isodibutylenes obviously belong, 
and the equilibrium between fructose, glucose, and mannose, when 
dissolved in dilute potash, affords another example of the same type, 
and one that is of special interest, as it involves the equilibrium of 
three dynamic isomerides. Keference may also be made here to 
the fact that whenever two optical isomerides become isodynamic, 
either alone or in presence of catalytic agents, an inactive mixture 
will always be formed. To the former class of substances belong 
all those which are incapable of being resolved into their optical 
isomerides by reason of the ** chaotic " nature of the molecule, whilst 
the use of catalytic agents has been illustrated by Holleman in 
the action of dilute potash on c^phenylglycollic acid {Bee. Trav. 
Chim.j 1898, 17, 323) and of acids and alkalis on the tartaric acids 
(t6u2., p. 66), the existence of an equilibrium being indicated in 
the latter case between the dextro-, Isbvo-, and meso-acids. The 
marked contrast observed between the behaviour of solutions of 
nitrocamphor when dissolved in chloroform alone, and when in presence 
of a trace of piperidine, serves to suggest the question as to whether 
a catalytic agent is not essential in all cases of isodynamic change ; 
in some instances, at least, this appears to be so, but in the fluid 
state the condition of a single dynamic isomeride must be regarded as 
essentially labile, since the isomeric change might be brought about 
by the action of a minute trace of alkali from the glass of the con- 
taining vessel, and it is possible that a trace of the second isomeride 
might prove sufficient ; in any case, the catalytic agent appears to be 
of the same order as the trace of moisture that is probably essential 
in all chemical changes, and may, perhaps, be regarded as supplying 
the conditions necessary for the production of a conducting circuit in 
the liquid. It is, however, of the utmost importance to notice that, 
whilst catalytic agents have a very great effect on the velocity of 
isomeric change, they do not produce any marked alteration in the 
eqtiilibrium which is ultimately established. The assumption made by 

Digitized by VjOOQIC 


Schiff (Ber., 1898, 31, 602) that ethylic acetoaoetate could be converted 
completely into the enolic or ketonic form by the addition of a trace of 
sodium ethoxide or of piperidine is obviously impossible, and has been 
completely disproved by the experimental work of Schaum (Ber., 
1898, 31, 1964), who has shown that the physical properties of the 
substance remain unchanged under these conditions; similar results 
have been obtained in the case of nitrocamphor, and from these it is 
obvious that, far from causing the complete conversion of the mixture 
into one of the dynamic isomerides, these agents are exactly those 
which would most rapidly bring about an equilibrium between the two 

A number of cases are known, and have been studied in detail, in 
which isomerides that are stable at ordinary temperatures become 
isodynamic when the temperature is raised ; the results obtained in 
these cases are of considerable importance in helping to solve the 
more difficult problems presented by substances which are isodynamic 
at ordinary temperatures. Thus, in the case of the stereoisomeric 
dibromotolanes, it has been shown by Wislicenus {DekaTuUsschrift, 
" TJeber die XJmsetzung stereoisomerer ungesiittigter organischer 
Yerbindungen ineinander bei hoherer Temperatur," Leipzig, 1890) that 
at 210^ an equilibrium exists in which the cia- and ^ran^-forms are 
present in the proportion of 48 : 52. 

(48 per cent.) ^»^^fl*^^ — ^aHfi'C-Br (52 per cent.). 

A similar equilibrium, although complicated by secondary actions, 
occurs in the conversion of ammonium thiocyanate into thiourea, 

NH^-ONS :::^ CS(NK^)^ 
(Volhard, J. pr. Chem,, 1892, [ii], 0, 11). The case which has been 
studied most fully is, however, that of the isomeric hexachloro- 
cycloketopentenes, described by Kiister in a paper, ** TJeber den 
Yerlauf einer umkehrbaren Beaktion erster Ordnung in homogenem 
System" (Zeit. physikal. CJi^m,, 1895, 18, 161); the equilibrium 
between these substances, which is reached in about 12 hours at 210^, 
corresponds with a mixture of 61*4 per cent, of the )S-cyclopentene 
with 38*6 per cent, of the y-compound, 

QQ\ nrj] G0\ CCl 

08=61-4 per cent.) J^j3^^j>C0 — y^j_^^j^CO(y«38-6percent.). 

In all these cases, equilibrium is reached only slowly at 210°, and 
at ordinary temperatures the isomerides appear to be stable ; they 
can, therefore, be separated by fractional crystallisation, or even by 
means of chemical agents; thus Kiister was able to estimate the 
quantity of the )8-compound in a mixture of the two isomerides by 

Digitized by VjOOQIC 


converting it into the sparingly solable anilide, OgClgO-NHPh, 
the y-compoiind remaining in solution unchanged. 

When the isomeric change proceeds with even moderate rapidity 
at a temperature of, say, 100% it will, as a rule, only be possible to 
separate the two isomerides by crystallisation, as chemical agents are 
found to have a great accelerating effect on the change ; the various 
modifications of ethylic diacetylsuccinate, studied by Knorr (AnnaHen^ 
1896, 293, 70), were separated by fractional crystallisation, and 
may be regarded as typical of this group ; so, too, the ketonic and 
enolic forms of ethylic anilidobenzylacetoacetate, described by Schiff 
{Ber.y 1898, 31, 205), can be fused several times, or crystallised 
repeatedly from spirit before equilibrium is reached between the two 

CHg-OO-CH-OOOEt _^ CH3-0(0H):C-000Et 


and may, therefore, be regarded as belonging to the same class of 
dynamic isomerides as the diacetylsuccinic ethers. 

Ti^Bromonitrocamphor is an example of a substance in which the 
isomeric change proceeds at a moderate rate even at ordinary tem- 
peratures, so that the two forms cannot be separated by the usual 
methods of fractional crystallisation. In this case, advantage is taken 
of the fact that the dynamic isomerides have nearly the same 
apparent solubility, so that, on allowing a solution in chloroform to 
evaporate to dryness, the two forms crystallise out side by side, and 
can be picked out mechanically. This appears to form almost the 
limiting case in which the dynamic isomerides can be separated, and 
when the velocity of change is greater, or the substances are liquids 
at ordinary temperatures, the task must be regarded as practicable 
only by the application of extraordinary methods. 

In some cases, it is possible to prepare the isodynamic forms of a 
substance by chemical methods, such as those employed by EUintzsch 
and Schultze, and by KonowaJoff in the case of the nitroparaffins (see 
p. 228), namely, conversion into the sodium salt and acidification with 
mineral acids. This was done by W. Wislicenus in the case of ethylic 
phenylformylacetate {Awruden, 1896, 291, 147), and the method has 
recently been applied successfully by Babe {Bw,j 1899, 32, 84) in the 
case of ethylic ethylidene- and benzylidene-acetoacetates, 


CHg-CH and CeH^-CH ; 


in the case of nitrocamphor, it was not found possible to prepare the 
pseudo-form in this way. In connection with the influence whicl^ 

Digitized by VjOOQIC 


bases and salts have been shown to have on the mutarotation of 
nitrocamphor, it is of interest to notice that, in order to separate the 
enolic forms described by Babe, it was necessary to use an excess of 
the mineral acid so as to neutralise, as far as possible, the accelerating 
effect of the bases in the solution ; this result is directly in accord- 
ance with the views put forward on p. 221 as to the action of these 
catalytic agents. 

The recent work of Schiff on ethylic acetoacetate is of interest as 
showing how dynamic isomerides may sometimes be prepared with the 
help of catalytic agents. In the condensation of benzalaniline with 
ethylic acetoacetate, he was able to show (Ber., 1898, 31, 601) that in 
presence of a trace of sodium ethozide only the enolic form of the 
additive compound was produced, whilst a trace of piperidine led to the 
production of the ketonic form. This action was assumed by Schiff 
to be due to the complete conversion of the ethylic acetoacetate by 
these catalytic agents into the enolic or ketonic form ; the incorrect- 
ness of this view has already been referred to, bub the influence of the 
two kinds of bases can be readily explained by means of equations 
similar to those which have already been used in the case of nitro- 
camphor, thus : 

COH _ CONa CisHijN 

or '6b. +NaOEt...M ^^^qj. _ 




C.C„H,3N + S0E* * • C.O„H, - + NaOEt. 




«H +C5HUN r: pNo,H„ c»»^»N 


The assumptions that are here made are (1) that the bases combine 
more readily with the ether than with its benzalaniline compound, and 
are, therefore, more effective in bringing about isomeric change in the 
former substance ; the reversed action in the above equations would 
then be only slight, as is indicated by the dotted arrows ; (2) that the 

Digitized by VjOOQIC 


benzalaniline interacts more readily with these basic compounds than 
with the free ether. 

On account of the equilibrium which exists in a mixture of dynamic 
isomerides, the addition to it of a chemical agent which interacts with 
only one form will necessarily bring about the complete conversion 
of the substance into that form. In this way, normal nitrocamphor is 
immediately converted by alkalis into a salt of the pseudo-form, 
whilst the anhydride of the latter is produced on heating it ; on the 
other hand, the salts are reconverted into normal nitrocamphor by 
acidifying a solution and crystallising the product, whilst the action 
of bromine and chlorine yields the normal bromonitro- and chloronitro- 
camphors described in Fart I. of this paper (Trans., 1898, 73, 986). 
For this reason, chemical methods cannot be used in studying sub- 
stances which are isodynamic at ordinary temperatures, and physical 
methods must be relied on almost entirely. This objection applies 
also to the use of agents, such as benzalaniline, which yield additive 
compounds of both types, and it is somewhat surprising that Schiff, after 
having demonstrated the enormous influence of a mere trace of sodium 
ethoxide on the interaction, should have been led to regard the forma- 
tion of the enolic derivative as proving that the ether was itself 
entirely enolic. It appears highly improbable that the enolic form 
of the ether should persist in a labile state, without undergoing 
isomeric change under the influence of the alkaline salts present 
during its preparation, or in the subsequent distillation, and such an 
assumption would only be justified after the nature of the ether had 
been definitely established by physical methods. 

The only chemical method that appears to be available for the study 
of dynamic isomerism consists in the use of ferric chloride as a colour- 
test for the presence of a hydroxylic modification ; in this case, how- 
ever, no product separates out from the solution, and as only a minute 
quantity of the substance is converted into ferric salt, the equilibrium 
between the isodynamic forms is not disturbed to any appreciable 
extent. The formation of the coloured ferric salt, and, therefore, the 
depth of the coloration, depend on the 'ratio of distribution' of 
the ferric oxide between the hydrochloric acid and the hydroxylic 
modification of the isodynamic substance, and as the latter is usually 
a very feeble acid, the amount of ferric salt formed will be exceedingly 
small compared with the quantity of ferric chloride added, unless the 
hydroxylic modification is present in large excess. The coloration will, 
therefore, give a rough indication of the extent to which the substance 
is present in the hydroxylic form, and in this way a reversible action 
may be made use of to study cases of reversible isomeric change in 
which unreversible actions cannot be employed. In the case of nitro- 
camphor and of ^-bromonitrocamphor, the test breaks down on account 

„.gitized by Google 


of the catalytic action of the ferric chloride, and this may be regai*ded 
as indicating a limit to the usefulness of the method. 

After a substance has been found to exhibit the dual chemical 
behaviour characteristic of isodynamic compounds, physical methods 
must be introduced to ascertain whether the case is really one of 
dynamic isomerism. The separation of two solid modifications of the 
substance is obviously of importance, since these may prove to be the 
two dynamic isomerides, but this must not be assumed to be the case 
without further evidence, as the differences of physical properties may 
be due simply to polymorphism. In order to establish completely the 
existence of dynamic isomerism, it is necessary to prove further that 
the isomerism persists in the liquid state^ although it is not essential 
for this purpose that the isomerides should themselves be isolated ; 
when, however, this can be done, the nature of the isomerism can be 
best established by examination of the recently-fused . substance, or of 
a freshly-prepared solution, which may show a gradual change of con- 
ductivity (Hantzsch and Schultze) freezing point (ibid.), refraction 
(Knorr), or rotatory power. In other cases, a substance may be proved 
to be a mixture of dynamic isomerides by means of variations in its 
physical properties depending on the treatment it has previously re- 
ceived ; thus the changes of density which occur in ethylic acetoacetate 
after it has been distilled (Schaum), may be regarded as affording 
proof that it is a mixture of isomerides or of polymerides, since the 
physical properties of a pure substance cannot be dependent on its 
past history, except in the case of solids. For many of the diketones, 
a proof of the presence of two or more isonftrides has been afforded 
by the determination of the magnetic rotation by Dr. Perkin (Trans., 
1892, 61, 801). 

Fdymorphiam, depending as it does on the different arrangement 
of similar molecules in the crystal structure, represents a form of 
isomerism which can only exist in the solid state and disappears 
immediately on dissolution or fusion ; but when, on the other hand, the 
differences in crystalline form are caused by differences in the 
structure of the molecule, the distinction between the various forms 
will persist for a longer or shorter time in the liquid state. The 
behaviour of the recently fused substance or of a freshly prepared 
solution affords, therefore, the best criterion in distinguishing poly- 
merism and dynamic isomerism on the one hand from polymorphism 
on the other ; the difficulty of observation would, of course, increase 
with the velocity of isomeric change, but even when this is too rapid 
to be detected, the constants of the liquid in the state of equilibrium 
would be those of a mixture of the two forms, and the composition 
of the substauoe might be established in a similar manner to that 

Digitized by VjOOQIC 


by whioh the constitution of the liquid diketones has been ascer- 
tained by Dr. Perkin from their magnetic rotations. 

The chief points of difference between the behaviour of isodynamic 
and of polymorphous substances are due to the time-factor which is 
always involved in connection with dynamic isomerides. In the 
crystallisation of polymorphous substances from solution, there is a 
sharp temperature limit above which one form only can separate in 
a stable condition, whilst below it another form is stable ; this 
limiting temperature represents the point at which the solubility 
curves of the two forms intersect, and it is not possible for them to 
crystallise out together from solution at any other temperature. If 
the temperature limit lies above the melting point of the substance 
or the boiling point of the solvent, one form will be known in a 
labile state only, and the addition to the solution of a crystal of the 
stable form will be sufficient to convert it completely into the latter. 
The behaviour of dynamic isomerides differs only from that just 
described in that the conversion of one form into the other by 
isomeric change requires a longer or shorter interval of time ; the 
limiting temperature will be that at which the ' apparent solubilities ' 
of the two forms are equal, but on account of the time required for 
the isomeric change to take place, it will be possible for the two 
forms to crystallise out together at temperatures considerably removed 
from this point, if the crystallisation is sufficiently rapid ; this 
behaviour has already been referred to in connection with the 
crystallisation of ir-bromonitrocamphor from chloroform solution, and 
is in itself sufficient to indicate that the two kinds of crystals differ 
in the structure, as well as in the arrangement, of the molecules. 

In dealing with the solid substances, similar results are obtained, 
the temperature of change being represented by the intersection of 
the vapour pressure curves of the two forms. But whereas the 
change of one form into another takes place at a sharp temperature 
limit in cases of polymorphism, there will be no such sharp temperature 
of transformation in cases of dynamic isomerism, unless the isomeric 
change takes place extremely rapidly. For this reason, the rapid 
change of crystalline form which is often observed when a substance 
is fused on a microscope slide and allowed to cool slowly may be 
regarded as specially characteristic of polymorphism. 

In conclusion, I desire to express my thanks to Dr. Armstrong for 
the interest he has taken in the work described in this paper, and 
for the help he has afforded in carrying it out. 

Gh£mioal Depabtment, 

City and Guilds of Londox Institute, 

Central Technical Collbos, South Kensington. 

Digitized by VjOOQIC 


XXVIL — The Action of Ammonia on Ethereal Salts of 
Organic Acids. 

By Siegfried Euhemann. 

Thb ethylic salts of glutaconic acid and its homologaes, as shown some 
time ago (Trans., 1893, 63, 259, 874 ; and 1898, 73, 350), when treated 
with ammonia, yield respectively aa'-dihydroxypyridine and the 
homologues of it which contain an alkyl radicle in the )8-position. 
Attempts to prepare the corresponding ^^substitution products of 
this dihydroxypyridine have already been made, but they were with- 
out result, as the ethereal salts which were required for the purpose 
could not be obtained by the interaction of the alkyl derivatives of 
chloroform and ethylic sodiomalonate (Buhemann, Ber,, 1896, 29, 
1016). Those pyridine compounds, as well as the derivatives of the 
dihydroxypyridine with alkyl groups in the y- and )9-positions are, how- 
ever, formed from the ethylic salts which are produced by the union of 
ethylic malonate and its homologues with the ethereal salts of the 
acids of the acetylene series (see Michael, J. pr, Chem,, 1887, [ii], 
35, 349; and Ruhemann and Cunnington, Trans., 1898,73, 1006). 

The ethylic salts of the acids containing an acetylene linking (I 
have as yet only used ethylic phenylpropiolate) combine with ethereal 
salts of fi-ketonic acids, such as ethylic acetoacetate and ethylic 
benzoylacetate, when the mixture is digested with sodium ethoxide. 
This reaction does not yield ethylic salts of unsaturated ketonic acids, 
but derivatives of a-pyrone, the union being accompanied by the loss 
of 1 mol. of alcohol. In using ethylic acetoacetate and ethylic 
phenylpropiolate, the reaction may be expressed by the following 
C-H^- C : C- COOaH, + CH^^^^^^s _ CeH^- C:CH- COOC2H5 

^ ^ ^ ^ ^^CO-CHj ■* COOOgHfi-CH-OO-CHs 

" COOO,H5-6:0(CH3)-^ + ^2^«^- 

The hope of transforming these compounds into derivatives of hydroxy- 
pyridine has, as yet, not been realised. 

This paper contains, besides the record of the work done in this 
direction up to the present, an account of some experiments which 
have been undertaken with the view of ascertaining whether the 
formation of imido-compounds takes place when ethereal salts of other 
saturated polycarboxyUc acids are acted on by ammonia, as in the 
case of propanetetracarboxylic acid (Ruhemann and Cunnington, loc* 

Digitized by VjOOQIC 


cit.). The constitution of the diamidoimide of this acid is, no doubt, 


.00— CH- OHj- CONH2 
represented by the formula NH<^q Ao-'CONH » ^^^ 

may expect that the ethylic salts of those acids which, like the former, 
contain carboxyl groups united with two neighbouring carbon 
atoms, will yield analogous substances, as with them the conditions 
are favourable for the production of a five-ring compound. The 
course of this reaction depends, however, not only on the position 
of the carbethoxyl groups in the molecule of the ethylic salt, but also 
on the nature of the other radicles present. Thus, ethylic )8-methyl- 
propanetetracarboxylate, when treated with ammonia, does not give 
an imidoamido-compound, but yields the tetramide of the acid. 
Further experiments with the view of investigating more closely this 
influence are in progress. 


Action of Ammonia on Ethereal Salts of Saturated Polyearboxylie 


EHiylic P-methylpropo/TietetracarboxyleUe, 


which is formed ^by the union of ethylic citraconate and ethylic 
malonate, is very slowly acted on by aqueous ammonia, and only a 
small quantity of the oil disappears after 6 to 8 weeks. The reaction, 
however, is accelerated, although it remains incomplete, by heating the 
ethylic salt with absolute alcoholic ammonia at 150^ for 2 days ; the 
amide, being sparingly soluble in absolute alcohol, separates out. It 
crystallises from boiling water in colourless prisms, which begin to 
decompose at about 255° and melt at 270° with evolution of gas. The 
following numbers were obtained on analysis. 

0*2162 gave 44*5 c.c. of nitrogen at 14° and 762 mm. N = 24-28. 
CgHj^N^O^ requires N = 24*34 per cent. 

In order to ascertain whether the above difference in the action of 
ammonia on ethyUc propanetetracarboxylate and its homologue is due 
to the use of the reagent in alcoholic solution and to the temperature, 
the experiment was repeated with ethylic propanetetracarboxylate 
under the same conditions. The solid which separated after heating 
the ethylic salt of propanetetraoarboxylic acid with an ezcesB 
of alcoholic ammonia for two days was found to be the diamido- 
imide of the acid. Its identity with the product obtained with 

Digitized by VjOOQIC 


aqueous ammonia was ascertained by a comparison of the physical 
properties and by a nitrogen determination. 

0-2032 gave 38-6 c.c. of nitrogen at 17° and 732 mm. N = 2M5. 
CyHgNgO^ requires N = 2111 per cent. 

Ethyl%cprop(mepentac€vrbox7/Iate, nH<^^'"9^'^^^ , which 

CO — Cn*OH(CONH2)2 
reacts with ammonia in the same way as ethylic propanetetracarb- 
ozylate does, yields the triamidoimide of the acid, the oil being 
completely dissolved when left in contact with concentrated aqueous 
ammonia at the ordinary temperature for 8 — 10 days. On concen- 
trating the ammoniacal liquor, crystals separate which readily dissolve 
in boiling water, and, on cooling, crystallise in colourless needles 
which melt and decompose at 212°. 
On analysis, it was found that 

0-1975 gave 40 c.c. of nitrogen at 20° and 759 mm. N = 2313. 
CgH^jN^Og requires N" = 2314 per cent. 

Ethylic phenylpropcmetriearhooeylate, ^6^6*^H'*^H(CONH-? ' ^^^^ 
treated with aqueous ammonia, yields, as was to be expected, the cor- 
responding triamide ; after 4 — 5 weeks, the oil disappears, and a solid 
separates which is very sparingly soluble in alcohol but readily in 
water, and crystallises from the latter in colourless needles which 
begin to decompose at about 200° and melt at 230° with evolution of 
gas. The ammoniacal mother liquor, on evaporation, yields a further 
crop of crystals of the same substance. On analysis, 

0-2035 gave 04320 CO2 and 0-1162 HjjO. = 5789 ; H = 6-34. 

0-2068 „ 30 c.a of nitrogen at 19° and 760 mm. N= 1670. 

CijjH^gNgO, requires C = 57*83 ; H = 602 ; N = 1686 per cent. 

Action of Ammonia on Ethereal Salts qf UnaatiJi/rated Folycarboxylic 
Adda, y-PhenyU aui-dihydroxypyridine. 

The starting point for the formation of this dihydroxypyridine isethylic 
phenylpropenetricarboxylate, the product of the union of the ethylic 
salts of phenylpropiolic and malonic acids. The formula 

^6^6 ^^H(COOC2H5)2 
indicates that this compound may be regarded as ethyUc phenyl- 
carboxyglutaconate, and that, under the influence of ammonia, it would 
condense to a pyridine derivative. This reaction takes place, yielding 
ethylic y-phenyl-aba'dihydroxypyriddne-p'Carboocylate. The action of con- 

Digitized by VjOOQIC 


centrated aqueous ammonia on ethylio phenylpropenetricarboxylate is 
very slow, only a small quantity of the oil being dissolved in the 
course of 2 — 3 weeks, but the action is more rapid with alcoholic 
ammonia at 100^ ; after several days, a solid separates, which is the 
ammonium derivative of the pyridine compound. This dissolves in 
water with the greatest ease, and the addition of hydrochloric acid 
precipitates a crystalline substance, which is only sparingly soluble 
in water but readily in ether or alcohol, and crystallises from dilute 
spirit in colourless prisms melting at 200°. Their alcoholic solution 
gives a purple coloration with ferric chloride. The result of the 

analysis agrees with the formula ^ ** 1* ^ ^ Tl 

^ ^ CeHg-C^iCH— C-OH 

0-2136 gave 05070 00^ and 0-0980 H2O. = 64-74 ; H = 5-09. 
0-2044 „ 10 c.c. of nitrogen at 20° and 754 mm. N = 5-55. 
C^^HigNO^ requires = 64-86 ; H = 501 ; N = 5-40 per cent. 

Ethylic phenylpropenetricarboxylate, on hydrolysis with baryta 
water, yields, as Michael (J, pr, Chem,, 1894, [ii], 49, 20) found, 
phenyglutaconic acid, and this, when saturated in alcoholic solution 
with hydrogen chloride, is transformed into ethylic'^ phenylgltUacanate ; 
after 24 hours standing, water is added to the solution, the oil which 
separates is extracted with ether, and the ethereal layer shaken with 
sodium carbonate. The substance which remains on evaporating the 
ether boils at 186 — 187° under a pressure of 11 mm. It is a colourless 
liquid, having a density d 18718°« 11017. The result of the analysis 

agrees with the formula C^Hj- CI OH -0000^^5^ 


0-2122 gave 0-5310 OOg and 0-1348 H^O. = 68-24^ H=7-05. 
OigHjgO^^requires = 6870 ; H = 6-87 per cent. 

The transformation of ethylic phenylglutaconate into the phenyl- 

^N , is best effected by heating 
it with concentrated aqueous ammonia in a closed tube at 100°. After 
2 days digestion, the oil completely disappears ; from the yellow solution 
thus obtained, hydrochloric acid precipitates a crystalline product which 
dissolves readily in hot glacial acetic acid, but with difficulty in boiling 
alcohol. From these solvents, it separates in colourless prisms which 
melt with decomposition at 254 — 255°. The yield of the dihydroxy- 
pyridine is very satisfactory, 5 grams of the recrystallised substance 
being obtained from 9 grams of the ethylic salt, corresponding with 
78 per cent, of the theoretical yield. 

Digitized by VjOOQIC 


The phenyldihydrozypyiidine was analysed, with the following 

0-1955 gave 05060 COg and 0-0855 H^O. = 7058 ; H = 4-85. 
0-2646 „ 17-6 CO. of nitrogen at 18^ and 755 mm, N = 7-54. 
OjiHgNOg requires = 70-59 ; H = 4-81 ; N = 7'48 per cent. 

The chemical behaviour of y^phenyldihydroxypyridine is analogous 
to that of the ^^-substituted aa'-dihydroxypyridines. In alcoholic 
solution, it gives a red coloration with ferric chloride ; dissolved in 
ammonia, it forms, on adding silver nitrate, a yellow silver compound 
which darkens on warming ; the ammoniacal solution turns green on 
exposure to the air, but the oxidation takes place more slowly than 
with the dihydroxypyridines which contain an alkyl group in the 
j9-position. It had been pointed out before (Ruhemann, Ber.^ 1893, 
26, 1559) that the j3-substituted aa'-dihydroxypyridines have various 
properties in common with resorcinol, one of them being the forma- 
tion of fluoresceins. y-Phenyldihydroxypyridine gives this reaction 
on fusion with phtbalic anhydride; on adding dilute potash to the 
product formed, a solution is obtained which is green in thin and 
red in thick layers ; the colour is discharged by acids. 


The homologues of ethylic malonate, as stated above, unite with 
ethylio phenylpropiolate with formation of ethylic salts of unsaturated 
acids, which may be transformed into aa'-dihydroxypyridines with 
alkyl radicles in the y- and /^-positions. 
„,,.«,,. T . . , 0-H.-C:OH-OOOOA 

— The reaction has been carried out with ethylic benzylmalonate, and 
the union of this salt with ethylic phenylpropiolate effected as in the 
former case. The mixture of the two compounds does not become warm 
on adding the sodium ethoxide, but if the mixture be heated at 100° 
for 12 hours, about one-third of the ethylic salts will enter into com- 
bination, and on again digesting the lower boiling portions of the oily 
product with sodium ethoxide, a satisfactory yield of the additive 
compound is obtained ; this is isolated in the usual manner. It boils 
at 260 — 265° under a pressure of 12 mm. and is a viscous, yellow oil 
having the density d 13713° = 1-1347. 

0-2005 gave 0-5202 00^ and 01220 B^O. = 70-75 ; H = 6-75. 
OjgHggO^ requires = 70-75 ; H = 6-60 per cent. 

On hydrolysis, this compound yields phenylbenzylglutaconic acid, 
the reaction being effected by digesting the ethylic salt with excess of 
alcoholic potash on the water-bath for 6 hours. The solid residue 

Digitized by VjOOQIC 


left on diBtilling ofiE the alcohol is dissolved in water and dilate sul- 
phuric acid added; effervescence takes place, and the organic acid 
is precipitated as a very viscous substance which does not solidify 
even after standing for some time ; it readily dissolves in alcohol and 
in ether. Without further purification, it is transformed into the 
ethylic salt by saturating its alcoholic solution with hydrogen chloride, 
and, after 24 hours, adding water to the solution ; the oil which 
separates is then dissolved in ether, the ethereal layer shaken with 
sodium carbonate, and, after removal of the ether, distilled under 
diminished pressure. 

n.Lri^n. 7 7. s C^Hg-CrOH-COOCaHg . 

J)tethyltcphenylbenzylgluiac<maie, c^H,- CH^ CH- 0000,4 ' " "" 

yellow oil which boils at 240 — 241° under a pressure of 10 mm. and 
has the density d 13713°= 1 '1082. 

On analysis, the following numbers were obtained. 

0-2037 gave 0-6598 CO, and 0-1 255 HgO. = 74-94 ; H = 6-84. 
O22H24O4 requires = 75-0 j H = 6*81 per cent. 

Monethylic phenyJhenzylgluiaconate is formed along with the former 
compound, and is extracted from the product of etherification of the 
acid by the treatment with sodium carbonate. On adding an excess 
of hydrochloric acid to the alkaline solution, a viscous, yellow oil is 
precipitated which partially solidifies in the course of a few days ; 
the solid product is collected with the aid of the pump, washed with 
very dilute spirit, and purified by recrystallisation from the same 
solvent. The colourless prisms thus obtained melt at 98° 

The analytical data agree with the formula 


O^jHg- CHg- CH-OOOH OeHg- OH,- CH- COOC3H5' 

0-2036 gave 05530 00^ and 0-1103 H^O. C- 7407 ; H = 6-02. 
01945 „ 0-5280 CO3 „ 01095 H^O. = 74-03; H = 625. 
C20H20O4 requires = 74-07 ; H = 617 per cent. 

The neatral solution of the monethylic salt in ammonia gives, 
with silver nitrate, a white precipitate of the corresponding stiver 
salt which is not acted on by light. It was analysed after being 
dried in a vacnum. 

0-2335 left, on ignition, 0*0585 Ag. Ag = 25-05. 

OjoHjgAgO^ requires Ag = 25-05 per cent. 

The transformation of diethylic phenylbenzylglutaconate into 
phenylbenzyldihydroxypyridine takes place very slowly, 5 — 6 days 
heating with aqueous ammonia in a closed tube at 150° being 
required before 4 grams of the oil entered into solution. The con- 

Digitized by VjOOQIC 


tents of the tube, the ammoniacal liquor, and a small qaantitj of a 
crystaUine product are freed from ammonia by heatiug in a vacuum ; 
hydrochloric acid is then added in excess, and the precipitate formed 
is washed with water and dilute alcohol. The substance crystallises 
from hot alcohol in faintly-coloured, lustrous plates which melt at 
176° to a brown liquid. 
The result of analysis agrees with the formula 


0-1862 gave 0-5325 OOg and 0-0937 H^O. C = 7799 ; H = 560. 

0-2338 „ 10-5 c.c. of nitrogen at 19° and 766 mm. N = 5-20. 

CigHigNOj requires C« 77-97 ; B[ = 5-41 ; N = 5-05 per cent. 

With a drop of ferric chloride, the alcoholic solution gives a purple 
coloration which is discharged by excess of the chloride. This 
dihydroxypyridine is oxidised under the same conditions as the 
others ; when its solution in alcohol is boiled, with free access of the 
air, it first turns green, and then brown. On adding silver nitrate 
to the substance dissolved in ammonia, a yellow precipitate is formed 
which darkens on warming. 

Farmaticn ofa-Fyrone DerivcUivee and their beltaviaur toioanrda 

The ease with which the union of ethylic malonate and its homo- 
logues with ethereal salts of acids containing an ethylene or acetylene 
linking is effected by sodium ethoxide induced me to study the action 
between the ethereal salts of acids of the acetylene series and those of 
j9-ketonic acids. This investigation has been undertaken especially 
with the view of transforming the compounds which were expected to 
be produced in this reaction into derivatives of hydroxypyridine, but, 
as mentioned in the introduction, a-pyrone compounds are formed 
which, under the influence of ammonia, do not yield pyridine deriva- 

Ethylic a'Pyrone-^^methyl-y-phanyl-P^'carhoxylate. 

The mixture of ethylic acetoacetate and phenylpropiolate does not 
evolve heat on the addition of dry sodium ethoxide ; the union of the 
two ethylic salts, however, may be effected by digesting the mixture 
on the water-bath for 5 — 6 hours, but, in order to obtain a satisfac- 
toiy yield of the a-pyrone compound, it is necessary to use a larger 
quantity of the ethoxide (about 2 — 3 grams) than in the preparation 
of the ethylic salts of the polycarboxylic acids. On treating the mix- 

Digitized by VjOOQIC 


ture with an excess of dilate sulphuric acid, crystals continue to sepa- 
rate for some time ; these are collected, washed with dilute alcohol^ 
and crystallised from hot alcohol, from which they are deposited as 
long, colourless needles melting at 104^. 

This substance is comparatively stable ; it passes over readily at 
207 — 214° under a pressure of 12 mm. with but slight decomposi- 
tion, as a yellowish oil which quickly solidifies. The pyrone com- 
pound is insoluble in water, but dissolves in ether as well as in alcohol, 
and these solutions do not give a coloration with ferric chloride. 

The analytical data agree with those required by the formula 


0-2013 gave 0-6128 CO^ and 0-1038 HgO. C = 69-47 ; H = 6-72. 
0-2024 „ 0-5165 OO2 „ 0-0997 HjO. = 69-69 ; H = 5-47. 
Cj^H^^O^ requires = 69-76 ; H«6-42 per cent. 

ffydrolysis of Eihylie Methylphenylpf/raneccMrboocylate, — ^The course 
which the hydrolysis of the compound takes is in accordance with the 
above-mentioned constitution. The reaction is accompanied by decom- 
position of the substance, with formation of phenylglutaconic and acetic 

acids, and is to be expressed by the equation ^ _?' Y* V]]>0 + 

^ ^ ^ 00002H5-C:0(OH3) 

3KH0 = ^«^^' 9-^^' ^^^^ + OHo- OOOK + O^HeO. 
CH2-000K ^ ^ * 

This transformation is effected by alcoholic potash ; on adding the 

latter to the ethylic salt, a red coloration is produced which disappears 

on boiling. After 4 hours digestion on the water-bath, the alcohol is 

distilled off, the residue dissolved in water, and hydrochloric acid 

added ; this precipitates a crystalline solid readily soluble in hot 

water, from which it crystallises in colourless prisms. The compound 

is characterised as phenylglutaconic acid by the melting point, 

164 — 166°, and by the results of analysis. 

0-2035 gave 0-4778 00^ and 0-0895 HgO. « 64-03 ; H = 488. 
Oi^HjoO^ requires = 64-07 ; H = 4-85 per cent. 

The presence of acetic acid in the products of hydrolysis of the pyrone 
compound was ascertained by the ordinary tests. 

Action of Ammonia on Ethylic Methylphenylpyroneccurboocylate, — ^The 
pyrone derivative dissolves in absolute alcoholic ammonia if left in 
contact with it for about an hour, and the pale yellow solution deposits 
a white, crystalline solid which rapidly increases in quantity ; this is 
collected on a filter, washed with absolute alcohol, and dried in a 
vacuum over sulphuric acid. The following analysis indicates thc^t 

Digitized by VjOOQ IC 


the pjrone compound has undergone a transformation which may 
be represented by the equation 

CeH^-CICH-CO^^^ 2 ^^ ^ CeHg-CICH- COONH, 
COOCjHs-CiqCHg)"'^ " COOC2H^-C:C(CH3)NH, ' 

0-2044 gave 0-4606 CO, and 0-1253 HjO. = 6145 ; H-6-81. 
0-2025 „ 0-4560 CO, „ 0-1265 H3O. C = 61-41 ; H = 6-94. 
0-2458 ,, 20 c.c. of nitrogen at 15° and 766 mm. N = 9'60. 
CijHjoNjO^ requires = 6 1 64 ; H = 684 ; N = 959 per cent. 

The compound thus produced is very soluble in water, and on heating 
at 100*", it evolves ammonia and turns yellow. It is characterised 
as an ammonium salt by the fact that, with silver nitrate, its aqueous 
solution yields a white, gelatinous precipitate of the corresponding 
silver salt ; this is not acted on by light, and although it turns brown 
when heated at 100°, it can be dried in a vacuum. 

0-2725 left, on ignition, 0-0767 Ag. Ag = 28-14. 
0-4090 gave 14 c.c. of nitrogen at 18° and 750 mm. N = 3-89. 
Oi5H:igAgN04 requires Ag = 2827 ; N = 3*66 per cent. 

On adding hydrochloric acid to the aqueous solution of the ammo- 
nium salt, a substance is formed which is not the expected pyridine 
derivative, but is ethylic methylphenylpyronecarboxylate. The acid, 
therefore, removes 2 mols. of ammonia from the ammonium salt and 
transfers it into the same pyrone compound from which, by the addition 
of ammonia, it was produced. 

Ethylic a-Pyrone-ay-diphenyl-pt'Carhoxylate. 

This is produced in the same manner as the former compound, by 
digesting a mixture of ethylic phenylpropiolate and benzoylacetate 
with sodium ethoxide on the water-bath for several hours, using in 
this case also about 2 — 3 grams of the ethoxide. The reaction is 
generally incomplete, a certain amount of the ethereal salts remaining 
unoombined; their presence often prevents the pyrone compound 
from crystallising out of the mixture, but the separation can easily be 
effected by fractionally distilling, under diminished pressure, the oil 
obtained on adding dilute sulphuric acid to the product of the reaction. 
First, the unchanged ethylic salts pass over, and then, at about 270°, 
under a pressure .of 12 mm., the pyrone derivative distils as a very 
viscous, yellowish oil with green fluorescence. This oil, after a 
short time, solidifies ; the solid readily dissolves in ether or alcohol, 
and crystallises from the latter in well-developed, colourless prisma 
which melt at 120—121°. 


Digitized by VjOOQIC 


The analytical results correspond with those required by the formula 


0-2021 gave 05553 00^ and 00910 H^O. C « 74-94 ; H = 500. 
CjoHjgO^ requires C = 75*0 ; H = 50 per cent. 

The yield of the pyrone compound is very satisfactory, as the 
unaltered ethylic salts which are contained in the lower fraction of 
the oil can be transformed into the pyrone by repeated digestion with 
sodium ethoxide. 

The further investigation of the pyrone derivatives described in 
this paper, and the study of the reactions between ethylic salts of 
other )3-ketonic acids and those of unsaturated acids of the acetylene 
series, is in progress, and will be published shortly. 



XXVIII. — The Changes of Volume due to Dilution of 
Aqueous Solutions. 

By Edwaed Brucb Herschel Wadb, M.A. 

Object of the Paper. 

The object of this paper is, first, to describe an apparatus intended 
for the study of volume changes consequent on diluting an aqueous 
solution ; secondly, to give an account of experiments performed with 
that apparatus. The experimental numbers obtained are not intended 
merely as illustrations of the applicability of the method (though they 
serve that purpose), but also as accurate values of the change of 
volume on dilution. The actual change studied in this paper is that 
which occurs when a certain volume of solution is diluted with <m 
equal volume of tvcUer, 

I. Review of Previous Work hearing on ike Suhfect, 

The method of attacking the problem hitherto invariably employed 
has been to determine the density of solutions ; and the researches 
carried out may be profitably subdivided into two classes, (a), those 
the direct object of which is the measurement of the contraction on 

Digitized by VjOOQIC 


dilutioii, (h), those in which the density was observed for ordinary 

Researches hearing directly on the subject of Contraction, — ^The first 
group includes the work of Nicol {Fhil. Mag,, 1883, [v], 16, 121), 
Oharpy {Compt. rend., 1889, 109, 299), and others. Nicol's plan was 
to determine the density of a solution of known strength, and then 
to dilute to exactly half that strength, and redetermine the density. 
His paper contains a table of contractions obtained in this way. 
Charpy's densities are carried out to four places of decimals, so that 
changes of volume less than 1 in 10,000 would escape notice. 
Marignac {Arch, Gen,, 1870, 39, 273) made observations of density to 
five places of decimals, in order to find the relation, if any, of con- 
traction to heat capacity. 

Other Researchesibearing on the Subject. — Determinations of density 
for various purposes have been made by many observers, but it will 
not be necessary to refer to those which are quoted to less than five 
places of decimals. Amongst those to be noticed are the results of 
Ctorlach {Zeit. anal. Chem., 8, 279), Marignac, Nicol, Pickering (Trans., 
1891), Bremer {Zeit. j^ysikal, Chem,, 1888, 3, 423), Kohlrausch and 
Hallwachs (Ann, Phys, Chem,, 1893, 50, 118 ; 1894, 63, 15), and, later, 
Kohlrausch {ibid,, 1895, 61, 185). 

Eocamination of escisting Density Determinations, — In most of the 
cases quoted, the first obstacle to the employment of the densities in the 
study of contractions is the way in which the results are stated. To 
calculate the contraction due to dilution from the data in an ordinary 
table of densities is a very serious task. Of course an actual error in 
arithmetic would not be excusable, but most of the tables of density 
are so constructed that one must have recourse to much interpolation 
in order to obtain the desired result. Again, when the very incon- 
venient calculation has been made, the results obtained by different 
authorities are not very concordant ; as an example, Table I (p. 256) 
gives the contractions of certain solutions as deduced from the tables 
of Nicol, Pickering, Marignac, and Kohlrausch and Hallwachs. The 
last two observers have utilised earlier observations by Gerlach 
and Kohlrausch to complete their tables. 

Explanation of TahU I. — »= gram equivalents per litre before dilu- 
tion ; n/2 = gram equivalents per litre after dilution ; F=> volume 
which holds 1 gram equivalent before; 2 F= volume which holds 
1 gram equivalent after dilution. The figure in the remainiug 
columns are the contractions in c.c. per 100,000 of the diluted 
solution as deduced from the following authorities. Nicol (N^) and 
(Nj) ; Kohlrausch and Hallwachs (K. and H.) ; Pickering (P) ; Marig- 
nac (M) (/oe. cit,). 

The discordance amongst the various authorities is perhaps not 

Digitized byCjOOQlC 


wade: the changes of volume dub to 

Table I. 
Sodium chloride. 






















































Sulphuric acid. 



























23 1 





































surprising if we remember that to establish one value of the contrac- 
tion by measurements of density involves the errors of four weighings, 
namely, (1) pyknometer empty, (2) full of water, (3) full of solution, 
(4) full of diluted solution. Neglecting further errors of temperature 
and concentration, the sum may very easily surpass one in 100,000. 
With regard to Kohlrausch and Hallwachs* method of determining 
the density, which claims a very high order of accuracy, it should be 
said that it is restricted to solutions whose density is below 1 -03, so 
that, for stronger solutions, they have adopted numbers obtained by 
less refined methods. 

II. The Experimental Method employed in the Paper, 

Features of ike Meikod, — The method employed will now be explained 
without dwelling as yet on details of procedure. The contraction of 
a solution on dilution with an equal volume of water is observed direct^ 

Digitized by VjOOQIC 







in an apparatus constructed specially for that purpose. Attention is 

fixed on the contraction, and no attempt is made to measure the 

absolute volumes very accurately. In order, however, to make the 

results obtained with different pieces of apparatus comparable and of 

general utility, the results are stated, not as actual contractions, but 

as contractions per 100,000 c.c. of the diluted solution, and for this 

reason some knowledge of the actual 

volumes mixed is, of course, necessary. Fio. 1. 

It will be shown, however, that a very 

imperfect knowledge of the absolute 

volumes suffices for this purpose. 

Descriptums of the Apparatus. — ^^ ||l|.~ \'i) 

Roughly, the apparatus resembles a 
U-tube whose two limbs are extended 
into bulbs of nearly equal capacity, 
BB', Fig. 1. These bulbs are provided 
with long capillary stems, CO', between 
which a scale is placed ; a few cubic 
centimetres of mercury, H, fill the 
bend of the tube. Thus the contents 
of the bulbs are isolated, as by a door. 
It was possible, as will be seen, to fill 
one bulb and stem with a solution, 
and the other with water. During 
the filling, the water and solution are 

kept entirely separate by the partition D ~ D' 

of mercury. To obtain a measure of 
the contraction due to mixing the 
water and solution, it is necessary first 
to set the apparatus upright in a 
water-bath and read the position of 
the menisci of water and solution on 
the capillaries; then to place the 
apparatus in a horizontal position, so 
that [the mercury barrier no longer 
intervenes and diffusion takes place, 

which may be promoted by slight 

shaking of the mercury; lastly, to 

set the apparatus upright once more and take a second reading of the 
menisci in the capillaries. The difference of the first and second 
readings is a measure of the contraction due to mixing. 

Accessory Apparatus. — Besides the actual contraction apparatus, the 
outfit consists of a water-bath, stirring-gear, reading telescope and 
mercurial thermometer. The water-bath was a galvanised iron tank 

_. , ^oogie 


holding 50 gallons, and the natural fluctuations of this large mass were 
exceedingly slow, so that it was easy to keep the water at constant 
temperature by placing a small gas jet under the tank and adjusting 
the height of the flame when necessary ; for such a large water-bath 
an automatic thermostat is quite superfluous. The stirring-gear 
consisted of two propellers driven by a large water motor ; for the 
construction of this very efficient apparatus, I have to thank Mr. F. 
Thomas, of Jesus Lane, Cambridge, who has had much experience 
in similar work. The thermometer was by Kicks, and gave actual 
elevations correct to 0*1° and, what was much more important, could 
be trusted to indicate differences of temperature in the water-bath to 
within about 0-005°. 

III. ThA Conduct of an Experiment, 

Filling ths Apparatus, — At the outset of an experiment, the 
apparatus has to be filled with water on one side and solution on the 
other. Some distilled water is well boiled to free it from air, and this 
water is drawn into the apparatus while hot ; to do this, one mouthy 
C, Fig. 2, of the apparatus is connected with the hot water in the 
funnel A. By alternately gently sucking and blowing at B, we can 
gradually fill one bulb with water free from air, without allowing it 
ever to transgress the partition of mercury. In exactly the same way, 
the remaining bulb can be filled with hot, boiled-out solution. 
Throughout the operation, the mercury separates water and solution. 
When the liquids are cool, the filling has to be completed, and pains 
must be taken that when the apparatus is vertical the two mercury 
surfaces are in the same horizontal plane, otherwise the ratio of the 
volumes of the liquids in the two bulbs will not be a definite one. 
The operation is referred to as '' levelling." 

Levelling. — ^This could be accomplished with the necessary accuracy 
by placing the apparatus vertically in a shallow glass trough, T, con- 
taining water. The water is removed by degrees, and if the observer 
looks upward obliquely, see Fig. 2, at the surface of the water, he sees 
the reflections of the mercury menisci descending to meet the real ones 
as the water is removed. If the two mercury menisci are in one 
horizontal plane, they will come in contact with their reflections at 
the same moment ; if they fail to do so, water is added to one bulb, or 
solution to the other, until the test is satisfied. The test was sensitive 
enough to show a change of 05 c.c. in either bulb, which was many 
times smaller than what was absolutely necessary. 

Readings. — We have next to obtain accurate readings of the menisci' 
on the scale of the capillaries. The menisci are brought into such a 
position that the horizontal cross-wire of the telescope appears to pass 

„.gitized by Google 



through them both, while the vertical cross-wire grazes the image of 
the scale ; the reading where the horizontal cross-wire intersects the 
scale is then recorded. To ensure accurate readings, we must observe 
the following precautions. 1. — The temperature of the water-bath 
must be kept steady, or else fluctuations must be allowed for, and kept 
within 0'02°. 2.— The apparatus must be exposed for a sufficient 
period to this temperature. 3. — Parallax must be avoided. With 

Fio. 2. 

regard to the temperature, it could be read with relative accuracy to 
about 0'005^y and owing to the large capacity of the bath it was 
usually constant to within about 0*02^ for a whole day.* 

To ensure that the apparatus had been exposed to this steady tem- 
perature for a sufficient period, readings were always continued until 
they became steady. With the very vigorous stirring employed, the 

* This is illustrated in Tables V and VII (pp. 264 and 266), which might be 
multiplied, and is of the greatest importance. The limit of accuracy is ultimately 
fixed by the steadiness of the water-bath. 

Digitized by VjOOQ IC 


steady state tisually sets in in about three-quarters of an hour, and 
then no further changes can be detected for six hours, except such as 
correspond accurately with the small and slow changes of the water- 
bath temperature. This point being settled in preliminary experi- 
ments, the apparatus was assumed to be in the steady state when its 
readings were unchanged during an hour and a half. Parallax was 
avoided by setting the capillaries vertical, and using a spirit level on 
the reading telescope. 

TJm Miaoing qf Water and Solution. — When the reading thus 
obtained has been entered in the notebook, short lengths of rubber 
tubing of narrow bore are slipped over the ends of the capillaries, and 
into these tubes are placed drawn-out pieces of glass tubing. The tips 
of the glass tubes are now sealed, and the apparatus is ready for the 
process of mixing. The best way to secure complete mixing of water 
and solution was found by preliminary experiments with coloured 
solutions, so that every stage could be watched. The apparatus is set 
horizontally with the bulb containing the salt solution uppermost. 
The denser solution descends owing to gravity, and almost without 
diffusion, so that two layers, one of water and the other of solution, 
can be distinguished in each bulb. A. slight shake of the mercury at 
this stage makes the mixing almost complete. The mereury is then 
transferred about 50 times from one bulb to the other, a slight 
shaking following each transfer. Further mixing will then not alter 
the reading obtained, and analysis shows that the concentration of the 
two bulbs is the same. (See section YI.) 

Second Reccing. — ^The reading, after mixing, is taken in just the same 
way as the first, the sealed ends of the glass tubes mentioned above 
being first cut off. The difference of the readings on the capillaries 
before and after mixing measures the contraction. 

Determination qf Concentration. — ^The concentration was in almost 
all cases found analytically in preference to working with synthetically 
prepared solutions; the objection to the latter course was that it 
would be necessary to dry out the apparatus before each experiment. 
The carbamide solutions, however, were made up to definite strength 
and the apparatus carefully dried, instead of using an analytical 
method. The chlorides were all titrated with silver nitrate, and the 
oxalic acid and potassium ferrocyanide by potassium permanganate. 
The strength of sugar solutions was found from their specific gravities 
with the help of Kohlrausch and Hallwachs' tables. 

lY. Standardisation of Appa/ratus and Sclutione. 

Th9 section devoted to the redaction of experiments (VI) will show 
that tha necessary data are, (1) the voluine of that part of the appa* 

^.giUzed by VjOOQ__ 



ratufi unoccupied by mercury, (2) the volume of a portion of the 
capillariee equal in length to one scale division ; from these, a factor 
(called the apparatus constant) can be obtained which, multiplied into 
the observed difEerence of the first and second readings, gives the con- 
traction in cubic centimetres per 100,000.* Besides this constant, we 
need to know the temperature coefficient of the apparatus, that is, the 
amount by which the reading of the apparatus is affected by a change 
of 1*0° and also a constant (called the inequality constant) which is 
deiined as the difference in volume of the two bulbs divided by their 
sum. The significance of this constant is explained in section TI. 

Determination qf the Volumes qf the Bulbs, — ^This was generally done 
by weighing the apparatus, first empty, then after the admission of the 
mercurial partition, and again after filling the remaining space with 
water. As a control method, the volume of water required to fill the 
space unoccupied by mercury was drawn in from a calibrated burette, 
and as another control the same space was completely filled with deci- 
normal sodium chloride, which was then washed out, by repeated 
washings, into a beaker and titrated with decinormal silver nitrate ; 
the volume of silver nitrate required is equal to the volume of liquid 
(other than mercury) which is contained by the apparatus. The two 
control methods were very useful, as after the apparatus was once 
mounted in its stand, it could not be conveniently displaced for 
weighing. The capillaries were standardised in the usual way by 
weighing columns of mercury occupying the greater part of their length 
and calculating their mean cross-section. The local cross-section was 
found by calibration with short columns of mercury. A preliminary 
calibration was always made, and if unsatisfactory the tube in question 

Tablb II. 


Apparatus constant. 





A. 1. 







(10-88 in experiments previous to October 
16th, 1897, owing to use of a larger 
mercury partition. ) 

Broken after only six experiments had been 

* The formula for this constant is 100,000 (;ir+ mi)/ ^a, where fir is the mass of 
mercury in length of right capillary equal to one mean scale unit, /ii the same for 
left capillary, p density of mercury. F the volume of the apparatus exclusive of 
that occupied by mercury partition. ^ . 

Digitized by VjOOQIC 



was rejected. All these standardisings were repeated at intervalB 
throughout the research, and a complete history of each apparatus was 
kept. Table II (p. 261) gives the constants of the pieces of apparatus 

The temperature coefficient of each apparatus was determined over 
a range of about 10° by reading the levels of the menisci at various 
steady temperatures. The following table will serve as an example. 

Table III. 
Temp, coeff. qf Tan. 3. 







19-96 Y 
19-97 j 






TJierrfore, change qf reading far 0-or= 0-012. 



Therefore, change qf reading for 0-01'*= 0-012. 

The inequality constant d was found by measuring the volume of 
each bulb separately in the way already described, and also by filling 
one bulb with a solution of sodium chloride and the other with water 
and mixing as in an ordinary experiment ; the loss in titre due to 
mixing gives a value of the inequality. The value of d can also be 
controlled by making a series of observations of contraction with 
the solution first in the right bulb ; secondly, in the left. If the 
proper value of d is used in the reduction of these results, the two 
curves will be identical (see YI.). 

Standardising the Volumetric Solutions. — ^The silver solution was 
prepared (1) by weighing the pure silver nitrate after exposure for 
10 minutes to a temperature of 120^, (2) by titrating against 
ammonium chloride specially resublimed just before use, (3) by 
weighing the silver chloride corresponding to a known mass of the 
solution, (4) by titrating with specially purified potassium bromide. 

„.gitized by Google 



The titrations for standardising were made both by weighing out the 
solntions, and also by using pipettes and burettes, the errors of which 
were accurately known. The permanganate solution was standardised 
by titration with oxalic acid and ammonium oxalate. When titrating 
the solutions of potassium ferrocyanide, the permanganate was 
standardised by means of a carefully recrystallised specimen of that 

Y. CcUcfdation qf the BesulU, 

The calculation of the results is, on the whole, simple. An 
example follows which it is hoped will make it clear. 

Tablb IV. 
Sodium chloride. 

































Column I gives the reading on the scale of the capillaries before 
mixing. Column II, the same after mixing. Column III, the 
difference. Column IV, the temperature of the water-bath at the 
time of recording reading in column I. Column V, the same for 
column n. Column VI, the difference (V-IV). Column VII 
shows the number in column III corrected for change of tempera- 
ture, to do which the number in column VI is multiplied by the 
eonstant a, and added to the number in oolumn III. Column VIII 
shows the contraction in cubic centimetres per 100,000 ; numbers in 
this oolumn are obtained by multiplying those in column VII by the 
constant L Column IX contains the volume of standard solution 
measured from a burette in order to titrate 10 c.c. (nominal) of the 
solution whose contraction is found in column VIII. Column X 
shows this volume corrected for calibration of the burette, and 
oolumn XI reduces it to gram equivalents per 100,000 c.c. of solu- 
tion. To obtain column XI multiply the figures in column X by 
constant 1 + c. Finally, column XII corrects for the inequality of 
the two bulbs. To obtain it, multiply column XI by l+d if the 
solution is in the right bulb, and divide by l+dii in the left. 

„.gitized by Google 



VI. Discussion of Errors, 

The errors attending each operation will be considered separately, 
and then the question of concordance of results will be raised. 

Possible Change of Volume qf the Apparatus. — ^A characteristic 
error in most volumetric work arises from changes in the volume of 
the measuring apparatus. In the method described, all the opera- 
tions can be carried out in a water-bath at constant temperature, 
and, moreover, a change of volume, even if it occurs, is only significant 
when it takes place between the initial and final readings. It is to 
be particularly noted that, as the apparatus is continually under 
observation during this interval, such a change could scarcely escape 
detection, and a long experience of the apparatus has not revealed 
any such change in the course of an experiment. If, however, a change 
occurred at the time of mixing, it could not be detected in the way 
described, and, therefore, to test the point, blank observations were 
made with water in each bulb. The example given shows that no 
change took place in the volume of the water, even when the 
apparatus was removed from the bath, and although the mixing was 
carried out much more energetically than in an ordinary experiment. 

Table V. 
BlamJc experimerU ; water in each bulb. 




Beading reduced to 
temp. 19-960^ 

• 1 2-80 

2 6-30 




Between 1 and 2, the apparatus was removed and mixing carried oat 

The difference in the last column is decidedly smaller than could 
be read with certainty, and corresponds to a change of only about 
0-02 in 100,000. 

Errors in Standardising. — Even great carelessness in standardising 
would scarcely affect the results, because the percentage error in the 
contraction is necessarily equal to the percentage error in the 
standardising by which it was deduced. The greatest contractions 
observed were, with very few exceptions, of the order of 40 units, 
that is, iO C.C. in 100,000 of the solution. An error of 1 per cent, in 
standardising would, therefore, cause an error of 0-4 such units, a 
quantity at most three or four times greater than the least which could 
be detected with certainty, and the error would be correspondingly less 

„.gitized by Google 



in smaller contractions. But it is altogether unlikely that the errors 
in standardising would amount to 1 per cent., as the independent 
determinations of the volume of the apparatus agree to within about 
0*3 per cent., and the standardising of the capillaries are more con- 
cordant than that. An example, taken from the standardising of 
Ton. 4, will illustrate this. 

Table VI. 

Standardisation of Ton, 4. 








February 17 




98 1 


98 1 


„ 20 

March 16 


„ 25 


„ 26 


V is the •* water volume " or part unoccupied by mercury, Vg the 
" mercury volume " or part occupied by mercury. The sum of these 
gives the whole volume (total calc.) ; this total may also be obtained 
by direct observation (total obs.). These totals ought to be constant 
to within the error of observation, but V and V^ are not necessarily 
the same in different fillings. After March 25, pains were taken to 
make Vg always equal to 6*5, otherwise the factor of the apparatus has 
to be recalculated at every filling ; /ir and /a; are the masses of mercury 
in a mean scale unit of the right and left capillaries respectively. 

The numbers in Table TI show that the errors due to standardising 
must be small, and the same conclusion may be drawn from the 
regularity of the results obtained in different appeiratus. See 
Tables X, XI. 

FarcUlactie Error* — ^This error is also insignificant. Thus ex- 
perience showed that if the telescope were put out of adjustment 
by a considerable distance, and the readings repeated, no sensible 
change was produced. 

Incomplete Mixing, — A possible source of error, and one which 
undoubtedly was present in the preliminary experiments, is imperfect 
mixing ; in all the final experiments, however, mixing was very 
thorough, and proved to be complete by analysis. In the example 
given below, a reading was taken, the contents of the apparatus were 
then mixed as in an average final experiment, and a second reading 
was taken 3 then a second very thorough mixing followed, and a 
third reading was taken. The second and third readings, reduced to 
a common temperature, are in agreement, showing that the first 
piixing was complete. 

Digitized by VjOOQIC 



Tablb VII. 
To iUustrate completeness qf mixing. 



Reading reduced to temp. 


28 00 

mixed aa in 
mixed again 

ordinary final experiment. 

as before. 


The error of inequality of tJie hvJhs appears to be completely avoided by 
the method of correction described in Y, which consists in increasing all 
the values of concentration in a certain ratio when the solution is in the 
smaller bulb, and diminishing them when it is in the greater. Great 
attention must be given, however, to the operation described as 
levelling (page 258), otherwise the ratio of the bulbs is not the same 
for the same apparatus in different experiments. The method of 
correcting was arrived at from a study of Kohlrausch's tables of 
density, which lead to the conclusion that if, say, V- 8 cubic centi- 
metres of normal sodium chloride are mixed with F+ S of water, the 
contraction is sensibly the same as if equal quantities of the two were 
mixed (provided 8 is small), the small deficiency of salt solution being 
almost exactly compensated for (as far as contraction is concerned) by 
the small excess of water. But the final concentration will, in one 
case, be below half normal, and in the other exactly half normal, for 
both the deficiency of solution and excess of water act the same waj 
on the final concentration. Hence the method of correction. The 
concordance of results by different pieces of apparatus with different 
inequalities, and using opposite bulbs to contain the salt solution, 
is considered to justify the correction. For instance, Table XI shows 
that, within the error of observation, sodium and potassium chlorides 
satisfy the equation X^n^/a where X is the contraction on 100,000 
cubic centimetres, n the number of gram equivalents in the same, and 
a, b, are constants. The concordance with this equation is not affected 
by the apparatus used, or the side of it which contains the solution. 

Errors in Titration, — ^In the greater part of the solutions examined, 
an error of 0*1 c.c. in the titration would correspond to a change in 
the contraction which could not be measured. 

Error of Temperature, — ^The most troublesome error is that due to 
measurement of temperature, and, unlike all the other sources of 
error, its effect is not proportional to the magnitude of contraction. 
The greatest percentage effects are, therefore, produced in the dilute 

Digitized by VjOOQIC 


solutions. In spite of the very thorough stirring and continued 
readings of the apparatus and thermometer, the uncertainty in the 
temperature of the apparatus (not the thermometer) may have reached 
0*01^. Two opposite errors of this magnitude would affect the 
apparent oontraction hy ahout 0-2 unit. 

Concordance of RcsvUm, — ^The general concordance of the final results 
leads to the conclusion that the error in the weakest solutions is 
about 0*2 unit, and that it may grow to 0*5 in the strongest.* Of 
ooorse this does not disprove the existence of a systematic error, but 
such an error, if it exists, would certainly tend to be exposed by using 
different patterns of apparatus. The five instruments employed varied 
widely in dimensions, yet no evidence of a characteristic error was 
found in them. 

In the weak solutions of sodium chloride, there is good concordance 
between the contractions measured directly and those calculated from 
Kohlrausch and Hallwachs' density tables. In the stronger, the con- 
cordance is not so good, but then in such solutions Kohlrausch and 
Hallwachs have availed themselves of earlier numbers by Gerlach, 
which are only stated to five places of decimals. A comparison is 
made in Table YIIL 

Table VTIL 

Compariacn of contractions (X) in cubic centimetres per 100,000 
observed by the author with those calculated from Kohlrausch and 
UaUvoaeht^ tables : nisihe number of gram equivalents per 100,000 
cubic centimetres qf diluted solution. 


10 2*0 X (Kohlrausch and 

Hallwachs) 2*8 

20 6-6 7*2 

80 180 147 

40 21*1 24-4 

60 30*9 82*5 

VII. Tlie Experiments. 

Material Employed. — ^The substances chosen for the measurement of 
contraction are representatives of the more important classes of com- 
pounds. Neutral compounds, which are supposed to remain un dis- 
sociated in solution, are represented by cane sugar and carbamide; 

* In the case of strong solutions of cane sugar and potassinm ferrocyanide, the error 
is somewhat greater, but in these solutions the method of finding the concentration 
was not so accurate as for the chlorides ; and, secondly, the contraction was increas- 
ing Tery rapidly with the concentration, so that errors in finding the concentration 
weze of greater importance than usuaL 

Digitized by VjOOQIC 



salts dissociating into two paris, by chlorides of sodium, lithiom, and 
potassium 3 salts dissociating into three, by chlorides of calcium and 
strontium; salts dissociating into more than three, by potassium ferro- 
cyanide. Mineral acids are represented by hydrochloric acid, and 
organic acids by oxalic acid. The purest commercial samples were 
used, and as far as possible they were recrystallised and their purity 
confirmed by analysis. Sodium chloride was recrystallised from strong 
hydrochloric acid, and the acid carefully removed, the salt washed with 
a little water, and finally ignited. The sample, on analysis, gave 
01 = 60*4 per cent. ; Calculated, 60*6 per cent. Potassium chloride on 
analysis gave 01 = 47*8 per cent.; Calculated 47 "6 per cent. A best 
commercial sample of lithium chloride was dissolved in water and 
filtered from the very slight trace of insoluble matter. The ratio of 
lithium to chlorine was then determined. Found 1 : 5*3 ; Calculated 
1 : 504. 

The observations marked* in Table XI were obtained with a com- 
mercially pure calcium chloride, the remainder with a solution on which 
special pains had been taken ; the latter was made by acting on per- 
fectly limpid crystals of Iceland spar with very pure hydrochloric acid 
made from redistilled sulphuric acid, and the sodium chloride already 
mentioned. The spar was in excess, and the solution was filtered 
from it. 

FrdimvMvry HxperimerUa. — Several series of preliminary experiments 
were made on sodium and potassium chlorides while learning the 
manipulation ; the earlier ones were faulty and discordant, but the 
later ones may be worth placing on record, although they did not 
attain the accuracy of the final series. 

Table IX. 
Preliminary ohaervcUians on KOI, NaCl. 
















n= number of gram equi- 







Talents per 100,000 cc„ 







and Xs= contraction in 







c.c. per 100,000 of the 







eolation due to 60 per 







cent, dilution. 

















Digitized by VjOOQIC 



The faalts known to be still present in these observations were 

(1) imperfect stirring, with consequent uncertainty of temperature, 

(2) insufficient attention to the ' levelling ' (see section III.). The 
first error was removed by the introduction of Mr. Thomas' stirring 
gear after the series on sodium chloride was completed and that on 
potassium chloride was about half complete. The second was not 
properly understood until the close of the two preliminary series, and 
from that time irregularities disappeared. 

27ie Final ExperimentaZ Results, — Tables X and XI give the 
results obtained when everything was in working order. X repre- 
sents as before the contraction in cubic centimetres per 100,000 and 
n the number of gram equivalents in the same ; X (calc.) are the 
numbers deduced from the equation ; X (calc.) = w*/a, which was found 
to answer well in most cases. 

Tablb X. 
KCl. a = 191; 6 = 1-64. 


X (calc.) 


















8 L 





18 16 



8 L 








8 L 












3 L 




3 L 

j 60-0 



4 L 




8 L 

1 62-45 



A 1 




8 R 




A 1 




4 L 




3 L 












8 L 

NaCl. = 250; h = 

















4 L 




8 L 




4 L 




3 L 




4 L 




8 L 








3 L 








3 L 












8 L 

Numerous observations were made on these salts in order to test 
experimental error and the influence of the apparatus used. The 
names of the appai*atu8 appear under T, and letters R, L, indicate 
whether solution was in right or left bulb. For the remaining 


Digitized by VjOOQIC 



Bubstances, a smaller number of observations have been thought 
sufficient, and the results are contained in Table XI, using the same 
notation as before. 

Table XI. 

a=12-59; 6=1-54. 

a=18-275; 6=1-52. 


JC (calc.) 



X (calc) 











40 . 






13 -4» 

CaM sugar. 





= 57-15; 6=2-00. 

! a= 

19-28; 6=1-528. 


X (calc.) 


1 71. 

X (calc.) 





























1 87-00 













= 29-45; 6 = 1-49. 



X (calc.) 
































1 55-99 


Digitized by VjOOQIC 



Table XL — carUintLed, 

K,FeC,N«. 1 




























VIII. Discfisaion of EestUts, 

Methods of Formulating the Resulta, — Several methods of represent- 
ing the results were attempted, and at first it seemed probable that 
an equation of the form X=An + Bn^ would be satisfactory, but as 
greater accuracy was attained it became clear that the deviation from 
this equation was systematic and greater than the error of obser- 
vation. Finally, the equation X^n^ja was used, and in most cases 
gives values of X agreeing well with observation. In the case of 
oxalic acid and potassium ferrocyanide, this equation does not, how. 
ever, seem to express the results. The comparison of the values of 
X calculated from the equation with those actually observed is made 
in Tables X and XI, and Fig. 3 (p. 272) shows curves drawn from the 
equation. The logarithms are plotted, not the actual values of X and n. 

Relation of Contraction to Equivalent Weight, — If the contractions 
due to diluting the various substances examined are written down in 
order of magnitude, they will be found to stand in the same order as 
the equivalent weights of the substances. This is shown in Tables 
XII and XIII, which contain the results for electrolytes and non- 
electrolytes respectively. 

Table XIL 

Name and equivalent weight of 

Contraction due to diluting from 
normal to half-normal. 

dissolved substance. 

























Digitized by 



Fio. 8. 









n=Ti'* of gram eq^iivaUnts in 
100,000 e.c, of solution aft&r 

X= contraction in e.c. per 
100,000 e.c. of solution, due to 
mixing equal volumes of water 
and solution. 

Table XIII. 

CO(NH,), 60 

Ci,H„On 342 


The curve, Fig. 4, shows the contraction as a function of the equiva- 
lent weight ; electrolytes alone are considered, and with the exceptioi) 

Digitized by VjOOQIC 



of potassium ferrocjanide the relation seems well marked. Non- 
electrolytes evidently contract very much less, other things being 
equal, than electrolytes, but there are not sufficient data as yet to con- 
struct a curve. 



. '^ 

^ - 

— c 





















Equiwahnt Weight. 

; — r-*i 1 








CorUniction Regarded ae Change in the Volume qf Dissolved Substance, 

Perhaps the most interesting way to look at the contractions is 
to consider them as changes in the apparent volume of the dissolved 
substance. For if the observed contraction is divided by n, the 
quotient is the change in the apparent volume of 1 gram equivalent 
of dissolved substance in undergoing 50 per cent, dilution. This is 
expressed in the equation 

^2n - ^ = ^h = n^-'^/a, 

where ^, ^ are the apparent volumes of 1 gram equivalent of dis- 
solved substance before and after the 50 per cent, dilution, n the 
number of gram equivalents per 100,000 after dilution (and, therefore, 
2n the number before). The solution of this equation is 

where a, b are known from the tables given on pp. 269, 270. The 
remaining constant can be found from one accurate determination of 
the density, and where such a determination has been placed on 

Digitized by VjOOQIC 



record it is possible to make complete tables of ^ at various 

Table XIV gives such values of ^. 

Table XIV. 
Substance NaCl. Temperature 18° 







Const, assumed. 






















17-97(0) ♦ 

This number is based on an observation by Gerlach. 

SvhaUmce SrClj. Temperature 18-0°. 













12 18(4) 












12-68 • 

This number is based on an observation by Eohlrausch. 

Substance CaClj. Temperature 18°. 

























This number is based on an observation by Eohlrausch. 

Digitized by VjOOQIC 



Table XIV. — continued. 
Substance HCl. Temperatttre IS"". 







Const, assumed. 





18 07(9) 

* Based on an observation by Eohlratiach. 
Substcmce O^jPSjs^ii* Temperature 18° 























* Based on an observation by Gerlach. 
Svletance KCl. Temperature 20°. 


























* Based on an observation by Nicol. 

Substance LiCl. 

Temperature 18° 


























Based on an observation by Kohlransch. 



The curves (Fig. 5) exhibit the values of ^. It should be remembered 
that the author is only directly responsible for the changes in ^^ and 
that the absolute values are affected by errors, if any, in the 
'* constant assumed" which is dependent on the accuracy of existing 
density determinations. 

The values of ^ are given partly in the hope that they will prove of 
practical value in physico-chemical work. In many branches of 
physical chemistry, solutions are prepared by dilution of stock 
solutions, and the experimenter wishes to know the concentration 
of the solution obtained, in, say, gram molecules per 100 grams 
of water. A continuous table of molecular volumes affords the 
easiest method of. arriving at the result. Partly, again, because 
the values of X only tell us the volume change on 50 per cent, 
dilution, whereas the tables of ^ gives us continuous information of 
the change of volume due to any given dilution. In the diagram, 
Fig. 5, ^ is shown as a function of n for various substances. As, 
however, it would be impossible to depict the various curves on the 
same page without greatly diminishing the scale, it has been neces- 
sary to subtract a constant round number from the values of ^ before 
plotting. These numbers are, for LiCl, 18; NaCl, 16; KCl, 27; 
HCl, 18 ; CaClj, 10 ; SrCla, 11 ; Cane sugar, 209. 


The following empirical laws have been arrived at by the experi- 
ments described. 

1. JT (as defined on p. 268, Table IX) is in almost all cases given by 
the following equation, X«^n^/a, and, consequently, ^ (as defined, p. 
273) must be given by 

ih = const. H ^^. ,, ■ 

^ a(2''-il) 

2. For comparable substances and concentrations, X increases with 
the equivalent weight of the substance in a regular way. 

It is hoped that these empirical rules, together with the tables of 
contraction and of molecular volumes, will be sufficient excuse for the 
publication of this paper. It should be added that the apparatus 
described is applicable without modification to many other physico- 
chemical problems, such as change of volume on neutralisation, change 
of volume on mixing two different liquids (for example, alcohol and 
water), changes of volume on formation of double salts in solution. 
The author hopes to proceed with the study of volume changes on 
neutralisation of organic acids without delay. 

The writer's best thanks are due to Prof. J. J. Thomson for 
permission to work in the Cavendish Laboratory, and to Mr. Griffiths 
and Mr. Whetham for the interest taken in the work throughput. _ 

Fig. 5. 





• •I 



/ / ^ 

10 20 30 40 SO 60 70 80 90 100 

Digitized by VjOOQIC 

278 baqnall: methanetrisulphonic acid, 

XXIX. — Methanetrisulphonic Acid. 

By Ebnbst Habold Bagnall, B.So. 

The sulphonic acids of methane, the Bimplest hydrocarbon, are Bub- 
Btances of more than ordinary intereet, and, although the mono-, di-, 
and tri-Bulphonic acids, CHj-HSOj, CH3(HS03)2, and CH(HS08)s, 
have long been known, no work on them has been published during 
the last fiye-and-twenty or thirty years. 

Of these, methanetrisulphonic acid, 0H(HS03)j,, the subject of this 
paper, was discovered by Theilkuhl {Annalen, 1868, 147, 134), and 
subsequently examined by Eathke {ihid.^ 1873, 167, 219). 

In attempting to prepare the various sulphonic acids and sulphones 
of dichlorobenzidine by the action of fuming sulphuric acid on dichloro- 
diacetylbenzidine, a substance was obtained which, on analysis, was 
found to be methanetrisulphonic acid ; as it can be easily prepared 
by the methods described in this paper, I have carefully examined the 
acid, and also studied the properties of its various salts. 

Preliminary experiments soon showed that the formation of methane- 
trisulphonic acid was really due to the presence of the acetyl groups 
in dichlorodiacetylbenzidine, as diacetylbenzidine under similar treat- 
ment gave a large quantity of the acid, whereas benzidine sulphate 
gave only benzidinesulphone and the various sulphonic acids of 
benzidine (Ber,^ 21, Rrf.^ 873 ; 22, 2467). Similarly, acetyl-a-naphthyl- 
amine gave a small quantity of methanetrisulphonic acid. 

As this reaction appeared to be a general one for the acetyl deri- 
vatives of benzidine and naphthylamine, it was anticipated that the 
acetyl derivative of aniline and the anilides of the higher fatty acids 
would yield methanetrisulphonic acid, or its homologues, as one of the 
products, and this was found to be so in the case of acetanilide. 
Acetanilide, when heated with fuming sulphuric acid, gives a remark- 
ably good yield of methanetrisulphonic acid, along with anilinedi- 
sulphonic acid, NH,- CeH8(HS08)2 [1:2:4], Other anilides were 
experimented with, but the results were not satisfactory ; the anilide 
of propionic acid gave a very small quantity of methanetrisulphonic 
acid, but neither ethanedisul phonic acid nor sulphopropionic acid could 
be detected (compare Buckton and Hofmann, J, Chnn, Soc.^ 1856, 0, 
p. 241). 

It is difficult to understand the general nature of these reactions. 
Much sulphurous anhydride and carbonic anhydride are always evolved, 
and it seems probable that some of the organic substance is oxidised to 
carbonic anhydride, and some converted into aromatic sulphonic acids, 

„.gitized by Google 


the acetjl group alone reacting with the fuming sulphuric acid to 
form methanetrisulphonic acid. Glacial acetic acid, however, when 
heated with fuming sulphuric acid, gives no methanetrisulphonic acid ; 
the sulphur trioxide dissolves in the acetic acid without giving off gas, 
and sulphaoetic acid, CH3(HS03)*COOH, is formed. It was antici- 
pated that, on heating acetamide with a large excess of fuming sulphuric 
acid, methanetrisulphonic acid would also be produced, but this is not 
the ease; the product consists entirely of methanedisulphonio acid, 
GU^iBSO^^ which was first isolated by Liebig (Annalen, 1835, 13, 
35), and subsequently prepared and examined by Wetherill (ibid., 1850, 
66, 122), Strecker {ibid,, 1856, 100, 199), and by Buckton and 
Hofmann (J, Chem. Soc.y 1856, 0, 241), and Rathke {Annalen, 1873, 
161, 152). 


Preparation of MethaaieiirisulpTionic Acid, 

Action of Fwning Sul/phuric Acid on Dichlorodiaeett/lbenzidine. — 
As stated in the introduction, it was from this substance I first 
obtained methanetrisulphonic acid. Dry powdered dichlorodiacetyl- 
benzidine (16*85 grams) was heated on a water-bath at 100^ with 
fuming sulphuric acid (100 grams, containing 70 per cent, free SO3), 
mixed with pure sulphuric acid (100 grams) until the whole dissolved.* 
The mixture was then heated for 3 hours on a sand-bath at a tempera- 
ture of 150^, with constant stirring, and left overnight. It was then 
poured on to ice, filtered to remove a dirty green precipitate which 
probably contains the sulphones and sulphonio acids of dichlorobenz- 
idine, and the dark brown filtrate neutralised with milk of lime, heated 
to boiling, and rapidly filtered from the precipitated calcium sulphate. 
To the clear filtrate, evaporated to a small bulk, sodium carbonate 
solution was added in quantity just sufficient to precipitate all the 
calcium as carbonate. This was filtered off, and the filtrate concentrated 
and allowed to cool slowly, when it deposited yellow crystals ; these 
were collected, dissolved in a small quantity of water, and glacial 
acetic acid added ; the white, crystalline precipitate of the sodium salt 
of methanetrisulphonic acid, Cn(!N'aSOg)g + SHjO, thus obtained, was 

* Great care is required in the mixing of fuming sulphuric acid with concentrated 
sulphuric acid ; to prevent accident, the concentrated sulphuric acid should be poured 
into the fuming acid, not vice versd. 

It is better to carry out these snlphonations in a porcelain beaker, the mouth of 
which IB covered with a close-6tting clock-glass. A hole is bored through the 
centre of the clock-glass, so as to admit a glass stirring rod which is connected 
with a water-motor. With this arrangement, very little sulphuric anhydride escapes 
into the air. 

Digitized by VjOOQIC 


0-1926, dried at 160°, lost 00276 H^O, Hfi = 14-33. 

0-2902 (anhydrous salt) gave 0-0363 COg and 0-1200 HgO. = 3*31 ; 

H = 0-45. 
0-1799 (anhydrous salt) gave 01188 'Nd^SO^. Na = 21-39. 
0-0785 „ „ „ 0-1720 BaSO^. 8 = 30-08. 

CH(NaS0s)8require8C = 3-72; H = 0-31 ; Na - 21-42,; S « 29-81 percent. 
CH(NaS03)8 + 3H2O requires HgO- 14-36. 

This salt crystallises in six-sided plates, with 3H2O, easily soluble 
in water, and is not, therefore, as suitable as the potassium salt for 
isolating methanetrisulphonic acid from an aqueous solution. On 
heating the anhydrous sodium salt in a tube, sulphur sublimes and 
sulphurous anhydride is evolved, together with a little hydrogen 
sulphide and carbon bisulphide ; the residue consists almost entirely of 
sodium sulphate. It is remarkable that when glacial acetic acid is 
added to an aqueous solution of sodium methanetrisulphonate, the salt 
is precipitated almost completely in crystals. 

Action of Fuming StUphuric Acid on Diacetylbenzidine, — In this 
experiment, 100 grams of fuming sulphuric acid containing 70 per 
cent, free SOg, 50 grams of concentrated sulphuric acid, and 20 grams 
of dry, powdered diacetylbenzidine were used, the operation being 
carried out exactly as in the first case, with the exception that the 
calcium salt was decomposed with potassium carbonate ; the potassium 
salt thus obtained is much less soluble in water than the sodium salt, 
and, therefore, more suitable for the isolation of the acid. The yield 
of potassium methanetrisulphonate obtained is very good, the above 
quantities yielding about 35 grams of the pure salt. 

Action of Fuming Sulphuric Acid on Acetyl-a-napWi,ylamine, — ^One 
hundred grams of dry, powdered acetyl-a-naphthylamine, prepared 
in the usual way, was heated with 100 grams of fuming sulphuric 
acid (70 per cent^SOg) and 50 grams of concentrated sulphuric acid, 
as in the other cases. The yield of methanetrisulphonic acid pro- 
duced in this way is but poor, considerable quantities of ammonium 
sulphate being formed during the reaction. 

Action of Fuming Sulphuric Add an Acetanilidc, — As this method 
gives the best yield of methanetrisulphonic acid, it is described in 
detail. Five grams of dry, powdered acetanilide were slowly added to 
a mixture of 30 grams of fuming sulphuric acid (70 per cent, free SOg) 
and 15 grams of pure concentrated sulphuric acid, the mixture being 
constantly stirred and heated at 130° for three hours. The amber- 
coloured, viscous product, when cold, was poured into water, boil ed, 
and neutralised with milk of lime ; not a trace of aniline was li be- 
rated. The liquid was filtered hot, and sufficient potassium carbonate 
solution added to the filtrate to convert the soluble calcium salts 

Digitized by VjOOQIC 


into the potassium salts. The filtrate from the calcium carbonate, 
when concentrated and left to cool, deposited beautiful, colourless 
prisms of almost pure potassium methanetrisulphonate. These were 
collected, drained on a porous plate, and recrystallised from boiling 
water. On analysis, they gave the following results. 

1 -7380, dried at 210°, lost 0-0803 HgO. HgO = 4-62. 

1-2043 (anhydrous salt) gave 0-1630 COj and 0-0346 HgO. C = 3-46 ; 

H = 0-31. 
0-5106 (anhydrous salt) gave 0*3600 KgSO^. K = 31-61. 
0-6259 „ „ „ 1-1972 BaSO^. 8 = 26-26. 

CH(KS08)3 requires C = 3-24 ; H = 0-27 ; K = 31-62 ; S = 25 94. 
CH(KS03)3 + H30 requires HjO = 4-64. 

This potassium salt loses its water of crystallisation at 210°, and 
when heated in a tube behaves like the sodium salt. It is insoluble in 
absolute alcohol, ether, and glacial acetic acid, and is only sparingly 
soluble in cold water, 1 part requiring 89 parts at 18° for solution ; it 
dissolves readily in hot water, from which it is again deposited, on 
slowly cooling, in beautiful, colourless prisms often over an inch in 
length. Phosphorus pentachloride has no action on the anhydrous 
salt. As in the case of the sodium salt, glacial acetic acid precipitates 
the potassium salt and all the other salts of methanetrisulphonic acid, 
unchanged from their aqueous solutions. 

To trace the part the aniline took in the action of fuming sulphuric 
acid on acetanilide, the mother liquor from the first crop of crystals 
of potassium methanetrisulphonate was further concentrated and 
placed on one side to cool, when a second crop of crystals of the tri- 
sulphonate was obtained ; these were filtered ofiE and the filtrate 
fractionally precipitated by glacial acetic acid. The precipitate, first 
formed on adding glacial acetic acid, was rejected as retaining gene- 
rally traces of potassium methanetrisulphonate. The further addition 
of glacial acetic acid to the filtrate produced a white, silky precipitate 
of the hydrogen potassium salt of anilinedisulphonic acid, 

NHj-CeH,(KS03)-HS08 [1:2: 4], 
first prepared by Buckton and Hofmann (loe. cit.) by the action of 
fuming sulphuric acid on aniline. The anhydrous salt gave the follow- 
ing results on analysis. 

0-3318 gave 0-3042 CO^ and 00662 H^O. C = 25-00; H«2-20. 
01807 „ 0-0541X2804. K= 13-43. 
0-2815 „ 0-4410 Ba804. 8 = 2151. 

0-3151 „ 13 c.c. nitrogen at 775-2 mm. and 20°. N = 4-92. 
NH5-CeH3(K803)-HS03 requires = 24-74; H = 2-06; 8 = 21-99; 
K = 13-40; N^4-81. 

1?o fiirther identify this substance, it was dissolved in water and 

Digitized by VjOOQIC 


treated with bromine water ; the yellow precipitate produced sublimed 
in beautiful, colourless needles which melted at 137^, and gave, on 
analysis, numbers corresponding with the formula of )3-tribromaniline 
GgHgBrg'NHs [1:2:4: 5]. The production of the acid potassium 
salt of anilinedisulphonic acid in the above reaction is due to the 
fact that glacial acetic acid, when added to an aqueous solution of the 
neutral potassium salt of anilinedisulphonic acid, throws down the 
acid potassium salt, which is much less soluble in water than the 
neutral salt. 

Action of Fuming Sulphwrio Acid on FropanUide, — ^Thirty grams of 
pure, powdered propanilide were added in small quantities to a mixture 
of 110 grams of fuming sulphuric acid (70 per cent. SO3) and 50 grams 
of concentrated sulphuric acid, and the whole heated at 110° for two 
hours ; on cooling, the product was poured into water, and, curiously 
enough, ethylic alcohol was given off. The solution was treated as in 
the other cases, first with milk of lime, and subsequently with a solution 
of potassium carbonate ; in this way, a substance was obtained which, 
after recrystallisation from dilute acetic acid, gave numbers, on analysis, 
showing it to be the hydrogen potassium salt of anilinedisulphonic 
acid, N'Hj-OeH3(KS03)-HS03 [1:2: 4]. By this method, anilinedisul- 
phonic acid alone is isolated; to obtain what little methanetrisulphonic 
acid is formed, the product was poured into water, neutralised with 
calcium carbonate, filtered, and the filtrate treated with excess of a 
solution of ammonium carbonate ; it was then again filtered, boiled with 
barium carbonate, and filtered while hot. On cooling, a small quantity 
of a crystalline barium salt separated which showed all the characteristic 
properties of barium methanetrisulphonate. The salt, dried at 180°, 
was analysed. 

01117 gave 0-0844 BaSO^. Ba = 4443. 

Ba3[CH(S03)3]2 requires Ba = 44-82 per cent. 

What part the propionic radicle takes in this reaction it is not easy 
to say 3 certainly a large quantity of the acid escapes during the re- 
action as free propionic acid, and furthermore, as just stateu, ctbylic 
alcohol is also produced, which must have come from this source. 
Neither ethanedisulphonic acid nor sulphopropionio acid could be de- 

MethanetristUphonic Acid and its Salts, 

Methanetrisulphonic Acid. — In preparing this acid, the pure barium 
salt was suspended in warm water and cautiously decomposed with pure 
sulphuric acid, a slight excess of the acid being used ; after heating 
on the water- bath for a few hours, the liquid was filtered from barium 
sulphate and evaporated ; the syrupy residue, when kept for 2 or 3 days 

Digitized by VjOOQIC 


in a good vacuum over sulphuric acid, deposited beautiful, colourless 
needles. The crystals, after draining on a porous plate in a vacuum, 
gave the following numbers on analysis, very accurate results being 
difficult to obtain as the acid is exceedingly hygroscopic. 

I. 0-6741 gave 00760 COg and 0-2001 H^O. C = 303; H = 3-29. 
n. 0-5480 „ 0-0557 COg „ 0-1636 HjjO. C = 2-77 ; H = 3-31. 

0-2720 „ 0-6025 BaSO^ ; S = 3042. 
CH(HS03)g requires C = 4-68 ; H = 1-56 ; S = 3760 per cent. 
CH(HS03)8 + 4H2O requires C = 3-66 j H = 3-65 ; S = 2926 per cent. 

This analysis appears to show that the acid has probably the formula 
OH(HS03)8 + 4H2O. Unfortunately, it is not possible to estimate the 
water of crystallisation by the ordinary method, as the acid decom- 
poses at 180^ before all its water of crystallisation has been given o£E. 

Methanetrisulphonic acid melts between 160° and 153°, and dissolves 
very easily in water with development of heat ; it is very deliques- 
cent^ has a pure acid taste, and turns blue litmus red. It is not 
changed, and no sulphuric acid is produced by boiling it with nitric 
acid, or when chlorine gas is passed through an aqueous solution of the 
acid. The salts of methanetrisulphonic acid are readily obtained by 
digesting the carbonates of the various metals with an aqueous 
solution of the acid. The barium salt is precipitated as a white, 
sparingly soluble, crystalline precipitate on adding barium chloride 
to an aqueous solution of the acid. The other salts are all more or 
less soluble in water, but insoluble in alcohol, and in glacial acetic acid. 
Phosphorus pentachloride has no action on it in the cold, but on 
slightly warming decomposition takes place ; it was not possible, how- 
ever, to isolate the product in a pure state. Methanetrisulphonic 
acid immediately unites with aniline with evolution of heat, forming 
the aniline salt, CH(HS03)3(CgH5-NH2V 

Silver Methcmetrisidp/wnate, CH(AgS03)3 + HgO. — ^To prepare this 
salt, freshly prepared silver carbonate was added- to a solution of the 
pure acid until the liquid was neutral, and the filtrate concentrated 
on a water-bath and allowed to cool, when slender crystals of the 
sUver salt gradually separated. The salt crystallises with IH^O, 
which it loses at 180°, but if it is heated above 180°, it darkens and 
begins to decompose. Silver methanetrisulphonate is soluble in water, 
but insoluble in absolute alcohol. 

2-3612 lost 0-0748 Kfi at 180° H^O = 3-18. 

1-2126 (anhydrous salt) gave 00915 COg and 01950 HgO. C = 205 ; 
H = 0-17. 

0-6477 (anhydrous salt) gave 0-3070 Ag. Ag = 6603. 

0-4298 „ „ „ 0-5104 BaSO^. S= 16-30. 

CH(S08Ag)8requiresC = 207; H = 0-17; Ag = 6613; S = 16-63 percent. 
CH:( AgS03)8 + HgO requires HgO = 3 -02 per cent. 

Digitized by VjOOQIC 


To determine the sulphur in the silver salt, Carius' method was 
first tried, but, as in the case of the calcium and barium salts, the 
silver salt crystallised out of the nitric acid quite unchanged at the 
end of the process, even after heating at a temperature of 210^; this 
is remarkable, as most silver salts of organic acids are decomposed by 
nitric acid under these conditions. The sulphur is easily deter- 
mined, however, by fusing the salt with pure potassium nitrate and 

It was anticipated that the ethylic salt of methsuietrisulphonic acid 
would be formed by heating silver methanetrisulphonate with ethylic 
iodide, and in order to investigate this point, 4 grams of the anhydrous 
silver salt were mixed, in a dry flask, with 6 grams of ethylic iodide 
dissolved in absolutely dry ether, and the mixture heated in a reflux 
apparatus for 6 hours; on extracting with ether and removing the 
ether by evaporation, no residue was left, and most of the silver salt 
was recovered unchanged. 

Barium Methanetrisulphonate, Ba3[CH(S03)8]2 + OHgO.— When a sol- 
ution of barium chloride is added to a solution of methanetrisulphonic 
acid or any of its salts, a most characteristic behaviour is noticed, and 
this may be used for detecting the presence of methanetrisulphonic acid ; 
on mixing the two solutions, nothing takes place for a few moments, 
and then, suddenly, a crystalline precipitate of the barium salt separates. 
The crystals were collected, well washed, and analysed. 

0-2189 (anhydrous salt) gave 0*1660 BaSO^. Ba = 44'68. 
0-5340 gave, at 220°, 0*0810 H^O. H20= 1516. 
Ba3[CH(S08)8j2 requires Ba = 44-82 and Ba3[CH(S03)3]2 + gH^O re- 
quires H20= 1501 per cent. 

Barium methanetrisulphonat-e dissolves very sparingly in boiling 
water, and crystallises on cooling either in the form of leaflets or 
needles, the temperature and concentration of the solution evidently 
determining the crystalline form. Strong nitric acid even under 
pressure does not decompose the barium salt, and the salt crystallises 
unchanged from 50 per cent, hydrochloric acid. 

Calcium Methamtrisulphanate, Ca8[CH(803)3]2 +I2H2O.— This salt 
crystallises with I2H2O in very small prisms which become anhydrous 
at 180°. It is very soluble in water, less soluble in dilute alcohol, 
and insoluble in absolute alcohol or glacial acetic acid, and remains 
unchanged when heated under pressure with nitric acid and potassium 

Capper Methamtrisulphanate, Cu3[CH(S08)3]2 + I2H2O.— This salt is 
obtained on decomposing barium methanetrisulphonate, suspended in 
boiling water, with the requisite quantity of a solution of copper 
sulphate, and the filtrate from the barium sulphate, after concen- 
tration on a water-bath, is left overnight over sulphuric acid in a 

Digitized by VjOOQIC 



Tacuam ; prismatic needles of the copper salt gradually separate, and 
these, after draining on a porous plate, were analysed. 

1-2784 lost 0-3010 H^O at 160^ HgO = 23-51. 
0-2954 „ 00780 CuO. Ou = 21-06. 
Cu3[OH(SOj)8]4+ I2H2O requires HjO = 2371 ; Cu=- 20-74 per cent. 

Ammonium Methaneirieulj^umate, OH(NH^S03)3. — This was pre- 
pared by neutralising an aqueous solution of the pure acid with 
ammonia, and exposing the solution in a vacuum over sulphuric acid, 
when crystals of the salt separated. 

0-2995 gave 00493 NHg. NH3 « 1646. 

0B:(NH4S03)g requires NHg- 1661 per cent. 

Ammonium methanetrisulphonate appears to crystallise in stumpy 
prisms or plates, probably belonging to the monosymmetric system. 
This research was carried out in the laboratories of the Owens 
College, Manchester, at the suggestion of Professor Perkin. 


For the accompanying measurements of silver methanetrisulphonate, 
and description of the crystallographical character of the other salts 
of methanetrisulphonic acid, I am indebted to Mr. W. J. Pope. 

Silver Methaneirieulphonaie. — ^This crystallises in small, white, opal- 
escent, iridiscent plates belonging to the orthorhombic system; the 
crystals are somewhat irregularly developed, and the faces in the zone 
[100:010] are deeply striated with lines parallel to the c axis. The 
extinction in a{100} is parallel to the edge ap, and the acute bisectrix 
emerges normally to a(lOO) ; the optic axial plane is c(OOl), the optic 
axial angle is large,and thedouble refractionis negative in sign and weak. 

Crystalline system. — Orthorhombic. 

a:h:e:^ 2-9152:1:0*5422. 
Forms present a {100}, b {010}, p {110}, g {011} and g' {012}. 

The following angular measurements were obtained. 


Number of 




ap =100:110 
bp =010:110 
pp =110:110 
pp =110:110 
bq =010:011 
qq =011:011 
qq" =011:012 
/^ = 012:012 


18'22'-- 19' 8' 
70 46 — 71 26 
14148—142 80 
87 81 — 88 11 
61 18 — 61 62 
66 81 — 67 28 
13 1—18 26 
80 14^ 80 89 

71 4 
142 9 
87 46 
66 69 
18 14 
80 26 


142 8 
37 62 

66 56 
80 22 


Digitized by Vo 




Barium SaU, — ^The crystals consist of minute^ rhomboidal shaped 
plates, the acute angle being 76^ j the extinction bisects the angles of 
the rhomb« 

Sodium Salt* — ^The crystals consist of long, six-sided plates, in which 
the extinction is parallel to one pair of edges ; the large face is per- 
pendicular to the positive bisectrix of a large optic axial angle. 

Potassium SaU. — ^The crystals consist of large, monosymmetrio prisms, 
showing the forms {100}, {001}, {110}, and {111}. The optio axial 
plane is perpendicular to the plane of symmetry, and one optic axis 
emerges nearly normally through the face (110). 

XXX. — MaltodextHn: its Oxidation- Products and 

By Horace T. Brown, LL.D., F.R.S., and J. H. Millar. 


I. Introduction .... 
II. Preparation of Maltodextrin 
III. Properties of Maltodextrin 
I Y. Oxidation of Maltodextrin . 
Y. Hydrolysis of Maltodextrinic Acid A . 
YI. Maltrodextrinio Acid B 
YIL Acid Hydrolysis of Maltodextrinic Acids A and B 
YIIL The Final Acid obtained by the Acid Hydrolysis 
of Maltodextrinic Acids A and B . 
IX. The Constitution of the Maltodextrinic Acids and 


X. Appendix, The Acid Hydrolysis of Starch 






I. Introduction, 

In a paper published thirteen years ago by one of us and G. H. 
Morris (Trans., 47, 1885, 527), a substance was described under the 
name of maltodextrin,* which had been isolated from a mixture of 

* The term maltodeztrin is dae in the first instanoe to Herzfeld, who, as far 
back as 1879, employed it to describe a derivatiye of starch which, in many respects, 
although not in all, corresponded with the subject of this paper (see Hezzfeld, 
Inaug.Di83.ffalU,lS79; also ^er., 1886, 18, 3469; ifttef., 1886, 19. 488). More recently, 
Armstrong suggested the term amyloln as being generally applicable to substances 
of this class. Both terms are oonyenient, since they recall the fact that the sub- 
stances in question, of which probably a series exists, may be regarded as combina- 
tions of amylin or dextrin groups, with amylon or maltose groups. ^ t 


the products obtained by the restricted hydrolysis of starch by malt 

' This was a non-crystallisable substance, having definite optical and 
reducing properties corresponding to [a]ys^ 193*1°, and k^^ 21*1. 
Expressed in the notation now usually adopted, these values are 
equivalent to [a]]> 180^. R 34'5, the reducing power of maltose 
being R 100. 

The homogeneity of maltodeztrin was proved by the persistence of 
the above optical and reducing powers under eontinued fractionation 
with alcohol, and also by the fact that it would pass slowly through 
a dialyser without change. Its unf ermentability with ordinary yeast, 
and the subsequently ascertained fact that it yields no crystallusable 
osazone with phenylhydrazine, proved the absence of ordinary maltose 
from maltodeztrin. 

One of the most marked characteristics of maltodextrin, and of the 
amyloms generally, was found to be the complete conversion into 
crystallisable and fermentable maltose under the action of the very 
active diastase of air-dried malt ; it was not converted into a definite 
mixture of maltose and a comparatively stable dextrin in the same 
way that soluble starch is under similar conditions. 

In the paper of 1886 above cited, it was considered that the facts 
known up to that time warranted the conclusion that the formula of 

maltodextrin might be written as \ ^^ xl^ ^"^ v ; that is, as a com- 

bination of two amylin or dextrin groups with one amylon or maltose 

Subsequent determinations of the molecular weight by the freezing 
point method appeared to confirm this view, but we do not lay much 
stress on this fact, since we are now convinced that the method of 
freezing is not by any means generally applicable to the non-crystal- 
lisable products of starch-hydrolysis. 

The present paper, together with the communications immediately 
following it, give an account of the principal results we have obtained 
during the past two years of continuous work on the subject of starch- 
hydrolysis. The latter part of the research has been carried on in 
the Davy-Faraday Research Laboratory of the Royal Institution, and 
we take this opportunity of tendering our thanks to the Directors 
and to Dr. A. Scott for the facilities they have given us. 

The methods hitherto employed for the separation and examination 
of the starch products have necessarily been very limited, and have 
for the most part been confined to (1) a determination of the optical 
and reducing constants of the products after as complete a separation 
as possible by alcoholic fractionation ; (2) an inquiry into the mode of 

Digitized by VjOOQIC 


action of diastase on the separate substances ; and (3) their behaviour 
with phenylhydrazine. Occasionally, attempts have been made to 
determine the molecular weight by the freezing point method, and 
sometimes also the action of yeast has been tried, with or without the 
addition of diastase. 

These methods, which are necessarily very imperfect, have often 
given different results in the hands of different experimenters ; indeed, 
so far is this the case that there may be said to be no unanimity even 
as regards the main facts of starch-hydrolysis. 

After turning our attention to the improvement of the existing 
methods of analysis of starch products, the results of which we laid 
before the Society in 1897 (Trans., 1897, 71, 72, 109, 115, 275), we 
endeavoured to supplement the older methods of attack with others. 

The one, which up to the present has been by far the most promising, 
is based on a study of certain well-defined oxidation products which 
can be obtained from some of the non-crystallisable starch derivatives. 
These are complex carboxylic acids which are capable of further 
hydrolysis by diastase and dilute acids respectively. An examination 
of the salts of these acids, and a determination of the exact amount of 
sugars which they yield on hydrolysis, has thrown much light on the 
constitution of the polysaccharide before oxidation, and the method 
appears to be capable of considerable extension. In the present 
paper, we shall confine ourselves to the study of the acid derivatives of 

II. FreparcUion of McUtodextrin, 

The preparation of maltodextrin in a state of purity, and in sufficient 
quantity to admit of satisfactory work, is, in itself, a difficult matter. 
It must be remembered that we have to deal with an absolutely 
unorystallisable substance, differing only very slightly in solubility 
from various other colloidal substances of a somewhat similar nature 
always accompanying it in the mixture of partially hydrolysed starch 
products from which it has to be separated ; it is only by long and 
very laborious fractionation with alcohol that the required result 
can be attained. If we commence with the mixed products derived 
from 2000 — 2500 grams of starch, we may consider ourselves fortunate 
if we can obtain 350 — 370 grams of crude maltodextrin, and this will 
probably be reduced to from 50 to 60 grams by the time it is sufficiently 
purified. Such a fractionation generally requires from two to three 
months to be effectually carried out. 

The mo<hu operandi which we have found to answer the best is as 

About 2 kilos, of well-washed potato starch are used in each opera- 

Digitized by VjOOQIC 


tion. Such an amount of starch, by the ordinary process of gelatinisa- 
tion, which must, of course, precede hydrolysis, would require as much 
as 50 litres of water, since a starch-paste of much more than 4 per 
cent, concentration becomes unmanageable. 

In order to reduce the volume of liquid to a reasonable point, so as 
to obtain the products of hydrolysis in a fairly concentrated state, the 
following device was adopted. Five hundred grams of the starch, that 
is, one-quarter of the whole amount, were gelatinised with 12 litres of 
water at 100° in the ordinary manner ; the starch-paste so obtained 
was then cooled to about 60°, and to this a trace of malt extract was 
added, just sufficient to destroy its viscidity. The solution was once 
more heated to 100° and made use of for gelatinising a further quantity 
of the starch, which was in turn rendered just liquid with malt extract 
after, of course, cooling to 60°. These operations were repeated until 
the whole of the 2000 grams of starch were obtained in solution in 
from 12 to 13 litres of liquid. This was finally hydrolysed to the 
required extent with an amount of finely divided malt equal in amount 
to about 2 per cent, by weight of the starch employed.* 

It is better to take a kiln-dried rather than an air-dried malt for 
this purpose, so as to ensure the limitation of the diastatic action. 
We prefer a malt with a diastatic power of about 38 according to 
lintner's standard. The temperature of the transformation should 
be 65°, and the progress of the reaction must be followed with the 
polanscope, and arrested by boiling when the specific rotatory power 
has fallen to about [ajo 168° tot 175°. 

In some cases, the filtered solution, containing about 15 percent, of 
mixed starch products, was submitted to fermentation for a few days 
with a little washed yeast, so as to remove the greater part of the 
maltose, but, on the whole, we prefer to separate the maltose by 
alcoholic fractionation only. 

The solution of mixed products, after evaporation to a thin syrup, 
is then poured into about 5 litres of alcohol of 90 per cent, (by 
volume), which is reduced to about 87 per cent, by the water of the 
syrup, and the mixture is digested on a water-bath for some hours at 

* We luiTe sometimes employed about 40 c.c. of a cold-water malt-eztract in- 
stead of malt. This is made by digesting finely divided malt with 2^ times its 
weight of cold water for 6 to 8 honrs. The filtered liquid must be heated to 65° 
ior about 15 minutes before it is used, so as to restrict the diastatic power. If this 
latter precaution is not taken, there is danger of running the hydrolytic process 
down too far, thus reducing the yield of maltodeztrin. 

t In the previous paper of 1885 on maltodeztrin, it was recommended to com- 
mence with a transformation whose united products has a specific rotatory power 
of about [a Id 180. We have since found that the lower value mentioned in the 
paper gives the best yield of maltodeztrin. 

Digitized by VjOOQIC 


the boiling point of the alcohol, with frequent agitation ; it is then 
allowed to cool slowly, and the clear alcoholic liquid poured off and 
distilled. The residue from the first digestion is now treated in a 
similar manner with successive quantities of 3 litres of alcohol of a 
uniform strength of 85 per cent., each digestion being continued for 
a whole day. In this manner, the original starch products are split 
up into 15 or 20 principal fractions, the early members of which 
contain most of the maltose, and have a lower specific rotatory power 
than the later fractions. It is the intermediate and later fractions 
which contain most of the maltodextrin, whilst the greater part of the 
' stable ' dextrin is left in the residue insoluble in 85 per cent, spirit. 

The first three or four of the principal fractions are rejected 
entirely, and the remainder, selected according to their specific 
rotatory power, are re-united to form two more parent fractions for 
subsequent treatment, one of them. A, being made up of 
fractions with an [a]i> of from about 164 — 173^, and the other, B, 
of all the remaining fractions possessing a specific rotatory power 
lying between about [a]© 174° and 182°. 

Fraction B, which may amount in such an experiment to 350 ot 
400 grams, consists for the most part of a mixture of various malto- 
dextrins with a little maltose which is detectable with phenyl- 
hydrazine. When treated with a little malt-extract (from air-dried 
malt) at 50° for an hour, it yields about 94 per cent, of maltose, so 
that the amount of stable dextrin in the mixture must be very small. 

Both A and B are further fractionated in an exactly similar 
manner, but using alcohol of 90 instead of 85 per cent. Selections 
from the resulting fractions are again made according to optical 
activity and reducing power, and those are again united whose pro- 
perties approach the maltodextrin sought, a similar process of treat- 
ment being continued until a product is obtained which is no longer 
capable of being further differentiated by alcohol. 

Such a procedure necessarily occupies a very long time, and the 
constant elimination of the less pure products results in a very small 
final yield of maltodextrin. 

We here give, in diagrammatic form, an example of one of many 
such fractionations which we have made during the progress of our 
work. In this particular case, which was not exceptional in any 
way, the complete separation occupied nearly three months, and 
entailed about sixty separations. 

Digitized by VjOOQIC 



-I I I 



o So <»*i * 

9 ^ » a ^ 















*j: I 





eo« 1 ^ 

I I 






«l I 

--I I 










CO r* . 

^ i-« 




**00 CO 


»«00 00 

"^ 00 Tl« 



--•s I 



00 00 eo 


« 1-1 

I— I 


Digitized byCjOOQlC 


III. Properties qf Maltodextrin. 

Maltodeztrin thus obtained appears to be a homogeneous product. 
It is a noD-crystallisable syrup which can only be completely dried in 
a vacuum over phosphoric anhydride in the apparatus we described in 
1897 (see Trans., 1897, 71, 76) ; in this condition, it is a colourless 
vitreous solid, somewhat hygroscopic, and soluble in hot water to an 
unlimited extent. It is soluble in alcohol of 85 per cent, to the 
extent of about 1 gram per 100 c.c, and is less sparingly soluble in 
alcphol of 90 per cent. 

It possesses the following optical and reducing properties.* 
[aJD 181—183° 
R 42—43 

The above specific rotatory power agrees very closely with the 

. observations described in 1885 {loc, eit,), but the reducing power is 

considerably higher than the value, R 34*5, assigned to maltodextrin 

in the earlier work. We are unable to account for this discrepancy, 

and must be content for the present merely to record it. 

Maltodeztrin, notwithstanding its high reducing power, certainly 
does not contain any maltose as such, since, unless previously hydro- 
lysed, it is not in 'the slightest degree fermentable with ordinary 
brewery yeast, and it does not yield a trace of crystaUisable osazone 
with phenylhydrazine. 

When a 5 or 10 per cent, aqueous solution of maltodextrin is 
digested for an hour at 50° with a few c.c. of malt-extract made 
from air-dried malt, hydrolysis proceeds very rapidly, and the pro- 
ducts of the reaction (due correction, of course, being made for the 
malt-extract) approximate closely in properties to those of pure 

It is only necessary to give the results of one such experiment, 
which is typical. A maltodextrin of [a]i> 181*1° and E 42*8, on 
hydrolysis under the above-mentioned conditions, gave a solution with 
[aJD 141*0° and R 94*0, the properties of pure maltose for a similar 
concentration being [ a ]d 138*0° R 100*0. Phenylhydrazine now gave 
a large yield of highly crystalline osazone, which, when separated from 
the trace of glucosazone due to the sugars of the malt-extract, proved 
to be maltosazone with its usual melting point. The crystals agreed 
in form with the well-crystallised osazone given by pure maltose, and 
they were unmixed with the " isomaltosazone " forms which in- 
variably occur when traces of dextrinous impurities are present (see 

* It will be noticed that these values are not consiBtent with the " law of 
relation " which we have prerionsly found to hold good between the optical and 
reducing properties of starch-prodacts. We shall have more to say on this point, 
and on the "law of relation," on a future occasion: 

Digitized by VjOOQIC 


Trans., 1895, 67, 731). These facts confirm us in our original opinion 
that maltodextrin is completely resolvable into maltose under the 
influence of diastase, and the want of complete accord of the optical 
and reducing properties of the hydrolysed maltodextrin with those 
of maltose is no doubt due to small errors of experiment incident 
to the hydrolytic process. 

When maltodextrin is hydrolysed by treatment with an ctoid, it is 
completely conrerted into d-glucose, whose properties were identified 
by crystallising the sugar and determining its optical and reducing 

Evidence that the final product of hydrolysis of maltodextrin under 
these conditions consists exclusively of dextrose was obtained in 
various ways, as shown by what follows. 

If maltodextrin is hydrolysed with oxalic acid^ and the excess of 
acid is removed with calcium carbonate, the optical and reducing con- 
stants of the solution are consistent only on the assumption that 
nothing but dextrose is present. 

Moreover, when such a solution of the hydrolysed products is 
firaeiumaUf//€nnerUed, it is found, on examining the solution at various 
intervals during the fermentation, that the optical and reducing pro- 
perties of the substance which has disappeared, and of that which is 
left at any stage, conform to the known optical and reducing factors 
of dextrose. 

In addition to this, we have carefully studied the quantitative 
relation of the dextrose formed to the amount of maltodextrin 
hydrolysed, in cases where the hydrolysis has been carried out with 
oxalic acid under the standard conditions fully described in the 
appendix to this communication. 

The dextrose was in these cases determined by two independent 
methods, (1) by the increase in specific gravity during hydrolysis, 
and (2) by the cupric reduction of the product. 

In this manner, it was found that 100 parts of maltodextrin, on 
complete acid hydrolysis, yield the following amounts of dextrose. 

1. — ^By increase of sp. gr. 108*7 dextrose. 

2. — By cupric reduction 106*0 „ 

A maltodextrin of the empirical composition represented by 

{/n^*w^n^\ fiJiould, theoretically, be capable of yielding 109 per 
cent, of dextrose. 

IV. Oxidation Of Afaltodextrin. 

The well marked reducing power of maltodextrin may be taken as 
evidence that its molecule contains the carbonyl group, and it there- 
fore seemed probable that, by limited oxidation, it would yield a 

„.gitized by Google 


hydroxy-acid or acida whcNse careful study might throw additional 
light on the constitution of the parent substance. 

The method which gave promise of the best results was that of 
oxidation with yellow mercuric oxide in the presence of baryta ; a 
process which, besides allowing of ready recovery of the products, 
admits of an estimation of the amount of oxygen used up in the 

A known weight of the maltodextrin was dissolved in water, and 
to the solution, which was kept hot on a boiling water bath, freshly 
precipitated mercuric oxide was added, keeping it always in con- 
siderable excess ; successive portions of a weighed amount of crys- 
tallised barium hydroxide were then slowly added so as to keep the 
solution just faintly alkaline. A portion of the liquid, filtered off 
from time to time, was tested with Fehling's solution, and the process 
was continued until no further induction was apparent. It was 
always found that, as soon as the power of cupric reduction dis- 
appeared, the further addition of b^yta to the hot neutral liquid 
caused a permanent alkalinity, and there was no further reduction of 
the mercuric oxide. The weight of the baryta added is a measure of 
the amount of oxygen used up in the oxidation, half an atom of barium 
corresponding to one atom of oxygen, or to the production of one 
carboxyl group in the mixed acids formed. 

When maUose is oxidised imder these conditions, we have always 
found that the end of the reaction with mercuric oxide corresponds 
almost exactly to the assumption of tkrte atoms of oxygen by each 
molecule of the sugar. 

On reference to a previous paper in which we have given careful 
determinations of the reducing power of maltose on Fehling's solution 
under standard conditions (Trans., 1897, 71, 100), it will be seen 
that the copper reduced also corresponds very exactly to the using 
up of three atoms of oxygen per molecule of maltose ; hence there is a 
probability that the course of oxidation of maltose is the same no 
matter whether we use Fehling's solution or mercuric oxide and 

The case, however, appears to be different when we come to 
tnaUodeaBtrin, If, in this instance, we estimate the apparent amount 
of maltose in the maltodextrin, with FehUng's solution, that is, if we 
assume the maltose constituent of the maltodextrin has the same 
reducing power on Fehling's solution as free maltose, then the amount 
of oxygen consumed when maltodextrin is oxidised with mercuric 
oxide and baryta is always greater than the calculated amount, in the 
proportion, of about 3 to 3*72. This difference may be due either to 
the oxidation of maltodextrin running a different course in the two 
processes, or to the fact that the maltose constituent of maltodextrin 

„.gitized by Google 


is not oxidised in the same manner by Fehling's solution as is free 

As an example of an experiment on the oxidation of maltodextrin 
with mercuric oxide and baryta, we may give the foUowing« 

84*28 grams of maltodextrin having a specific rotatory power of 
[ajo 181 1 and a cupric-reducing power of R 42-8 were dissolved in 
1000 c.c. of water, and the solution, after being heated in a porcelain 
evaporating dish on the water-bath, was treated with mercuric oxide 
and barium hydroxide in the manner already described. The point at 
which further reduction ceased was very sharply marked, and corre- 
sponded to the consumption of 62 grams of Ba(0H)2 + SH^O, or 3'72 
atoms of oxygen for every molecule of apparent maltose in the original 
maltodextrin, as determined from the value of R. 

The decomposition of the barium salts and the complete separation 
of the barium, present certain special difficulties. Attempts to effect 
this by accurate precipitation with sulphuric acid are only partially 
successful, owing, in the first place, to a slight solubility of the barium 
sulphate in the dextrinous liquid, and, secondly, to the fact that a 
certain amount of the precipitate is in such an extremely finely 
divided state that it cannot be removed by ordinary filtration, and 
consequently it is impossible to accurately hit the point of complete 
decomposition of the salts. 

After many trials, we have found the following method the best for 
surmounting these difficulties. 

The solution of the oxidised products, after filtration, is cooled down, 
and a stream of carbon dioxide is passed through it in order to 
neutralise the very small excess of baryta present.* The solution of 
the barium salts of the organic acids is then evaporated to a thin 
syrup and the greater part of the barium is thrown out by adding 
dilute sulphuric acid. 

The solution of acids containing BaSO^ in suspension is then placed 
in a centrifugal machine, and the perfectly clear liquid, still, however, 
containing some barium, is poured off when complete separation has 
taken place, and is again evaporated to a fairly thick syrup. To this, 
about 1 C.C. of strong hydrochloric acid is added, and then a fair excess 
of sulphuric acid, the solution being kept cold. Under these condi- 
tions, the barium is completely thrown out of solution. A mixture of 
95 per cent, alcohol containing some ethylic acetate is at once added. 
This precipitates the dextrin-like acids, which are then dissolved in 
a small quantity of water and again thrown out with alcohol and 
ethylic acetate. This process is repeated until all traces of mineral 

* In some cases, the solation at this point was treated with hydrogen sulphide to 
remove any traces of mercury which might have passed into solation. As a matter 
of fiict, however, no mercoric snlphide was ever obtained. ^ . 

Digitized by VjOOQIC 


acids are removed, when the aqueous solution can be finally freed from 
the last traces of barium sulphate in the centrifugal machine. 

The fractionation of the acids can then be proceeded with, but as 
they are much more soluble than the maltodeztrin from which they 
are derived, fractionation with 95 per cent, alcohol is insufficient. 
Advantage can, however, be taken of their insolubility in ethylic 
acetate, and of the fact that they show a differential solubility in 
mixtures of ethylic alcohol and the acetate. The method we have 
usually employed is treatment of the crude acids with successive 
portions of hot alcohol by the process of digestion employed in the 
fractionation of the starch products, and then precipitation of the 
alcoholic solutions with ethylic acetate. By continuing this treatment 
through 20 — 30 fractionations, it was found possible to separate the 
deztrin-like acids derived from the oxidation -of maltodextrin into two 
series, one, the more soluble portion, having a specific rotatory 
power of [a]o 176° — 179°, and the other with a rotatory power of 
[ajo 189 — 192°. It IS this second main fraction with the higher 
rotatory power, and constituting the larger amount of product, which 
we have submitted to detailed examination. 

When in a pure state, it is an uncrystallisable, dextrin-like substance, 
having a specific rotatory power of [a]D 1923°, and without the 
slightest reducing action on Fehling's solution. For convenience, we 
will provisionally refer to this acid as maltodextrinic acid A, 

It has distinctly acid properties, and forms a well-defined, uncrystal- 
lisable calcium salt, which can be prepared by boiling a solution of the 
pure acid with calcium carbonate, and precipitating the salt with 

The calcium salt contains 2*4 per cent. Oa. 

The great interest attached to this acid derivative of maltodextrin 
lies in the fact that it is hydrolysable both by diastase and by acids, 
giving, with the enzyme, a definite amount of matton and an acid of 
lower molecular weight, whilst by acid hydrolysis it yields a definite 
amount of dextrose and a final acid of still less complexity. 

Y. Hydrolysis of Maltodextrinic Acid A with Diastase. 

Under the ordinary conditions of hydrolysis with diastase, this 
acid yields maltose and an acid of lower molecular weight, which we 
shall refer to provisionally as maUodeoBirinic acid B, 

The reaction was studied quantitatively in the following manner. 

A solution of the A acid, containing about 5 grams per 100 cc. 
was digested for 2 hours at 50° with 5 cc. per 100 of a cold water 
malt extract made from air-dried malt. The specific rotatory power 
rapidly fell from [a]i> 192*3° to [a]o 162°, at which it remained 

Digitized by VjOOQIC 


constant,* the reducing power simultaneously rising from E. to 
E. 40*0 (maltose =100). The solution gave a pure maltosazone on 
treatment with phenylhydrazine, and a further proof of the nature 
of the reducing substance was obtained by submitting the hydrolysed 
solution to fermentation with a little washed and pressed yeast. 
When the fermentation was complete, the optical activity and 
reducing power of the solution were redetermined, and from the 
differences between these and the original values two distinct esti- 
mations of the sugar fermented [were made, on the assumption 
that it was maltose. The results were as follows. 

Grams of maltose 
fermented per 100 c.c. 

Estimated from optical activity 2*290 

„ „ reducing power 2416 

The concordance of these separate determinations leaves no doubt 
that the reducing substance produced by the diastasic hydrolysis of 
the A maltrodextrinic acid is maltose. The A acid has, therefore, 
yielded 40 per cent, of maltose and 60 per cent, of some other 

VI. MaUodextrtnic Acid B. 

This can be isolated from the products of the diastase hydrolysis of 
the A acid in the following manner. 

The mixed products are fermented as completely as possible with 
a little washed yeast, and the residue, which is decidedly acid in 
character and amounts to 60 per cent, of the original A acid, is 
boiled with a little calcium carbonate. After filtration, the solution is 
evaporated, treated with dilute alcohol to remove a trace of proteids 
derived from the yeast, and the calcium salt is precipitated by more 
alcohol ; this is further purified by re-dissolving it in water several 
times and throwing out by alcohol. At this point, the calcium salt 
was found to contain 3*1 per cent, of calcium, but as it was 
not yet quite pure, it was submitted to still further treatment. 
The salt is in the first place re-converted into the free acid by 
exctct precipitation with oxalic acid. To the solution, filtered from 
calcium oxalate, tribasic lead acetate is added, and after separating a 
small precipitate which is formed, the solution is treated with hydrogen 
sulphide. The filtrate from the lead sulphide containing the acid is 
precipitated with strong alcohol, and the precipitated maltodextrinic 
acid B is repeatedly evaporated on the water-bath with alcohol until 
the acetic acid is completely removed. The acid is again con- 
It is, of course, understood that the nanal corrections were made for the malt 


Digitized by VjOOQIC 


verted into its calcium salt by boiling with calcium carbonate^ and 
the salt is separated by alcohol precipitation. 

It is a white, amorphous powder which^ after drying at 100*, 
contains 3*8 per cent, of calcium. 

YII. Acid Hydrolysia of Maltodextrvnic Acids A and B. 

Both A and B maltodeztrinic acids yield definite amounts of 
dextrose on hydrolysis with acid, with the separation of a final residual 
acid which is identical in the two cases. 

These reactions have been studied quantitatively by acting on a 
definite amount of the acids, or their salts, with oxalic acid, under 
the standard conditions described in the appendix. The amount of 
dextrose obtained from the free A acid in this manner was found 
in one case to be 85-8 per cent., and in a second 85*3 per cent. The 
calcium gait of the B acid, under similar conditions, gave 67*7 per cent, 
of dextrose. 

The sugar, separated from the hydrolysed liquid by crystallisation, 
proved to be ^glucose, without the admixture of any other sugar.* 

VIIL The Final Acid obtained by the Acid Hydrolysis of Maltodextrinic 

Acids A and B. 

This acid is, with simultaneous production of dextrose, the final 
result of the acid hydrolysis of both the A and the B acids. We have 
obtained it from both sources, but it is of course more convenient to 
prepare it by direct acid hydrolysis of the A acid, without any 
intermediate use of diastase. It can be isolated from the products 
of oxalic acid hydrolysis by two methods. By the first, the solution, 
after neutralisation with calcium carbonate, is, in the first place, 
fermented to remove the dextrose, and is then evaporated and treated 
with strong alcohol to precipitate the calcium salt. By the second 
method, fermentation is avoided, the neutralised liquid, after evapora- 
tion^ being treated with stnmg alcohol for the removal of the dextrose. 
On the whole^ we prefer to omit fermentation. 

The calcium salt thus obtained is further purified by precipitation 
from its aqueous solution with alcohol of 95 per cent, with the 
addition, as occasion demands, of a little ether. When alcohol alone 
is used, the salt is apt to come down very imperfectly and in such a very 
finely divided state that it remains suspended, forming a jnilky liquid. 

* In one of these experiments on the B calciam salt, the oalciom oxalate which 
separated during hydrolysis was estimated. Its calciam was foand to correspond to 
8*8 per cent, of the salt taken, a result which agrees exactly with the direct de- 
termination of calciam in the original B salt. 

Digitized by VjOOQIC 


This milk can, however, be made to ** break " by shaking with a little 
ether. Where the object is purity of product rather than yield, the 
addition of the ether should be as far as possible limited. 
. The first preparation of the calcium salt of the final acid from the 
hydrolysis of maltodeztrinic add A gave 10*1 per cent, of calcium. 
By hydrolysis of the B acid, the salt was obtained also containing 
lO'l per cent, of calcium. In the former case, the fermentation 
method had been employed. 

By very careful and repeated fractionation of the calcium salt, a 
product was finally obtained containing 10*44 per cent, of calcium, but 
beyond this it was found impossible to go. 

A combustion of this calcium salt with potassium dichromate in a 
stream of oxygen and air gave the following results. 

0-2148 gave 0-2666 COj and 01004 H2O. C = 324 ; H = 51. 
{G^^fi^\C2k requires « 324 ; H = 4*9 ; Ca = 108 per cent. 

The study of this acid has been much facilitated by the discovery 
that the same final acid is present in considerable quantity amongst 
the products of oxidation of malta$e by mercuric oxide and baryta. 
The oxidation of maltose was carried out in exactly the same manner 
as that of maltodextrin already described. The calcium salts of the 
acids obtained were hydrolysed with oxalic acid,* and the final acid 
was purified as a calcium salt in exactly the same manner as the final 
acid from the hydrolysed maltodextrinic acids. 

In this state, two preparations gave, after drying at 100% percent* 
ages of calcium equal to 10*4 and 10*5 respectively. 

The latter preparation, on combustion with dichromate, gave the 
following results. 

I. 0-2636 gave 0-3023 COg and 01161 HgO. 0-324 ; H«50. 
11.0*2131 „ 0-2621 COj „ 00964 HjO. = 32-2; H = 60. 
(05HgO^)30a requires « 32 4 ; H = 4-9 ; Oa = 10'8 per cent. 

There can be no question as to the identity of the final acids ob- 
tained by the oxidation and hydrolysis of maltodextrin and maltose 
respectively, as both their properties and their percentage composition 
are similar. 

The acid itself, when liberated from its calcium salt, is a colourless 
syrup which shows no tendency to form a crystallised lactone. It is 
very soluble in water, but almost insoluble in alcohol of 96 per cent. 

* Daring this hydrolysis, considerable quantities of dextrose were liberated. The 
oxidation of maltose under the above conditicms evidently results in the production 
of biose acids which are resolvable on hydrolysis into C«- and Cp-groups. We are 
sabmittiBg this interesting fact to dose examination. 

Digitized by VjOOQIC 


Its cinchosic salt shows no tendency to crystaUise^ but the acid 
appears to form a orystallisable hydrazide, the only crystalline 
derivative we have been able, np to the present, to prepare. The 
caloinm salt is very soluble in water, but quite uncrystallisable from 
that solvent ; it has only a very slight influence on polarised light, 
and the actual amount of rotation is still doubtful. 

The great obstacle to the extended examination of this acid has 
hitherto been the difficulty of its preparation in sufficient quantity ; 
now, however, that its identity with the acid similarly obtained from 
maltose is put beyond all doubt, we shall be able to prepare it in con- 
siderable quantities and submit it to a careful study. 

Taking into consideration the method of its formation, its source, 
and the empirical formula of its calcium salt, it would appear to be a 
normal carboxylic acid derived from a pentose. It has certain pro- 
perties in common with xylonic acid produced by the oxidation of 
xylose, but our detailed study of the acid in question is at present 
too incomplete for us to speak positively on this point. It appears to 
us, however, that the preparation of a Gg-derivative from a C|2-<sarbo- 
hydrate in the manner we have described is an interesting fact, and 
that in the limited oxidation and hydrolysis of polysaccharides of C^- 
sngars we may have an approach to an explanation of the formation 
of pentose derivatives by living plants. 

In connection with this subject, we are also reminded of O'SuUivan's 
work on the arabinic acids (Trans., 1884, 46, 41 ; 1890, 67, 59 ; 
1891, 60, 1029), in which he showed that arabic acid under the 
hydrolysing influence of dilute sulphuric acid yielded a series of acids 
of gradually diminishing molecular weight with simultaneous produc- 
tion of arabinose and in some cases galactose. 

IX. Constitution of the MaUodexirinic Adda and Maltodextrvn. 

From its mode of preparation, and its behaviour under successive 
hydrolysis with diastase and oxalic acid, it is clear that maltodex- 
trinic acid J[, the main product when maltodextrin is oxidised with 
mercuric oxide and baryta, must be a carboxylic acid derived from a 
polysaccharide. Since the original polysaccharide, maltodextrin, is a 
reducing substance completely hydrolysable to maltose by diastase, 
it must be regarded as being made up of a certain number of C^g or 
maltose residues with the elimination of the elements of water, but 
still retaining, like maltose itself, a terminal COH group. If the 
process of oxidation of the polysaccharide consisted merely in the 
transformation of this terminal GOH into carboxyl, we should expect 
to obtain an acid with its terminal C^^'fi^^P ^^^ ^P ^^ ^^^ malto- 

Digitized by VjOOQIC 



bionic acid type.* Such an acid, when treated with oxalic acid, would, 
like maltobionic acid, be transformed into glucose and gluconic acid, 
whilst under the action of diastase it would be almost certain to yield 
maltose and free maltobionic acid. 

Although it is very probable that under the oxidising influence of 
bromine^ maltodextrin may ultimately yield a complex acid of this 
constitution, it is evident that when mercuric oxide and baryta are 
used the oxidation proceeds much farther, and results in the splitting 
up of the terminal Cj^^-group into a C^-acid residue which remains 
attached to the hydrolysable portion of the polysaccharide, and 
adds of low molecular weight which are separable by the sub- 
sequent treatment to which the complex acid is submitted during 

It has been shown by Scheibler and Mittelmeier (Ber,, 1890, 23, 
3060; ibid., 1893, 26, 2930) that the formation of reducing poly- 
saccharides, that is, of compound saccharides with an aldose or ketonic 
structure of the terminal group, may be explained by the union, with 
elimination of the elements of water, of the carbonyl of one of the 
constituent groups with an alcohol group of another. 

If we accept Fischer's constitutional formula for maltose as 


\ O / 

a polysaccharide corresponding to the empirical formula < >, ^no * 

which we have hitherto employed for maltodextrin, will be repre- 
sented by the union of three molecules of maltose, in the manner 
described above, with the elimination of 2 mols. of water. The 
constitutional formula of maltodextrin may then be represented as 














* Maltobionic acid, OuHgOj^ was obtained by E. FiBcher and Hcyer by the 
oxidation of maltose with bromine {Ser., 1889, 22, 1943). From the fact that it 
splits np, nnder the inflaence of dilnte mineral add, into <i-glacose and <i-glaoonio 


,.gitized by 


This may be written in an abbreviated form thns, 

The <[ IB used, as suggested bj Scheibler and Mittelmeieri to 
denote the open terminal COH-groap. In this scheme, the (\f- 
groups are represented as linked to one another through oxygen 
atoms, and it is at these points that the elements of 2 mols. of water 
are introduced during diastase hydrolysis, with the formation of 
3 mols. of maltose. 

When the hydrolytic agent is an acid, the complex molecule 
behaves as though it were made up of O^^-groups, also linked through 
the intermediary of an oxygen atom, and, just as in the former 
case, it is at these points that the elements of water are introduced, 
with the consequent production of glucose. Maltodextrin should 
therefore take up the elements of 6 mols. of water in order to be 
completely converted into glucose, 100 parts of maltodextrin yielding 
109 of dextrose. How near this conclusion corresponds to the facts 
has already been shown (p. 293). 

It may be objected that, according to the above constitution of 
maltodextrin, there seems to be no sufficient reason why diastase and 
acids should not both hydrolyse it to glucose, since, constitutionally 
considered, there is no apparent difEerence between the linkings of the 
C^-and 0^12*^^^?^* 

There may, however, be a difference of spatial relations between 
the adjacent Oj^^-groups which may be sufficient to determine the 
differential action of the diastase, as contrasted with the absence of 
that action on the constituent Cg-groups, where we have every 
reason to believe the configuration is identical* 

Another possible explanation may be found in the relative size or 
mass of the G^- and Cig'S^^^P^' which may have some effect on the 
result. We have probably an instance of this kind in the selective 
effect qt the enzyme glucase on a-methyl-><2-glucoside on the one hand, 
and a-methyl-xyloside on the other, as shown by Fischer (Z0U. 
phynol. Chem., 1898, 26, 68). 

Of these two glucosides, one derived from a C^- and the other from 
a Cg-sugar, and built up on an exactly similar plan, both constitu- 

acld, the concltuiion was drawn that maltobionio aoid is derired from maltose by 
the oxidation of the terminal COH-groap. 

We have onraelyes prepared maltobionio acid in the above manner, and hare 
studied its products of hydrolysis qoantitatively. These results which will be 
found described in the appendix to this paper, are fully confirmatoiy of the oon- 
elusions of Fischer and Meyer. 

Digitized by VjOOQIC 


tionallj and stereochemicallj, the derivative of ^glucose is readily 
attacked by glucase, whilst that of xylose resists the action of the 
enzyme. The same difference also exists between the action of emulsin 
on the )9-methyl-c{-glucoside, and the )3-methyl-xyloside respectively. 

Possibly an investigation which we are making on the influence of 
the two enzymes, glucase and diastase, on maltodextrin may deter- 
mine which of the above explanations is probable. 

If the oxidation of maltodextrin resulted merely in the conversion 
of the GOH-group into carboxyl, an acid would be formed the con- 
stitution of which, written in the abbreviated form, would be re- 
presented as follows. 

This would be the acid already referred to, which would almost 
certainly split up into maltose and maltobionic acid under the action 
of diastase, and into c^glucose and ^gluconic acid when hydrolysed 
with oxalic acid. 

The constitution of maltodextrinic acid A certainly cannot be 
represented in this manner, for it yields a C^-acid on complete 
hydrolysis, not gluconic acid. The only explanation possible is that, 
daring oxidation with mercuric oxide and baryta, the terminal 0^2" 
group is attacked in such a manner as to remove, by simultaneous 
or successive oxidations, (\B.^fi^f leaving the residue of a CgH^^OQ 
acid still in combination with the polysaccharide residue. The only 
possible constitution of the maltodextrinic acid A, and the only one 
which, as we shall presently see, is in consonance with all the known 
facts, is, the following, 


The reducing C^j'^^^P shown in the extended constitutional 
formula of maltodeictrin has in fact now become 


which is the resi(kie of a normal carboxylic acid derived from a pentose. 
When the maltodextrinic acid A is hydrolysed with diastase (I), and 
with oxalic acid (II), the reactions take the following course. 

Maltodextrinic Maltodextrinic Maltose. 

acid A, acid B, 

Digitized by 



n. fC:0,,^0, + 4H,0 = 40aH,,0, + CsHioO.. 

Malt«deztrinic cf-Qlacose. Pentose acid, 

acid A. 

In a Bimilar manner^ the B acid is also split up into glucose and the 
pentose acid under the action of oxalic acid, with the assumption of 
only 2 mols of water. 

A careful examination of the facts which we have recorded in the 
experimental part of this paper will show that they are fully in accord 
with the above generalisations, as regards the composition, constitu- 
tion, and reactions of the derivative acids. This is brought out more 
clearly in the following summary. 

A Acid, — The calcium found in the purified salt was 2*4 per cent., 
which agrees exactly with the percentage of calcium in a salt of the 
composition {0^^4902q)20&, the empirical formula for the salt of 
maltodextrinic acid A. 

On hydrolysis with diasUue, the A acid gave, in two experiments, 
40*0 per cent, of maltose. The reaction I requires 42 per cent. 

On complete hydrolysis with oxalic acid, the A acid yielded, in one 
experiment, 85*8 per cent, of glucose, and in another 85*3 per cent. 
The reaction II requires 88*4 per cent. 

jB Acid, — ^The calcium found in the pure salt was, in two separate 
preparations, 3*8 per cent. A salt of the composition (Oi7H^Oj0)2Ca, 
the empirical formula of the B salt, should contain 3*9 per cent, of 

The B calcium salt, on complete hydrolysis with oxalic acid, yielded 
67*7 per cent, of glucose. Its theoretical yield is 70*7 per cent. 

As to the empirical formula of the final C5 acid, obtained by the 
complete hydrolysis of the maltodextrinic acids, there can be no doubt ; 
but further investigation is necessary before we can positively state 
whether it is, as its mode of preparation would suggest, a normal 
carboxylic acid of a pentose. 

The evidence as to the constitution of maltodextrin afforded by a 
study of its oxidation products seems to us conclusive, and it is further 
strengthened by the fact of the complete hydrolysis of maltodextrin 
to maltose under the influence of diastase, and by its yielding, within 
the experimental limits of error, the theoretical amount of glucose on 
treatment with oxalic acid. 

This further study of the subject has served only to strengthen 
our previous views of the constitution of maltodextrin, although we 
must admit that some of the older evidence, such as that derived from 
the relation of optical activity to reducing power, and the molecular 
weight determinations, was not such as we should now deem conclusive. 

Digitized by VjOOQIC 


We may still, if we choose, regard inaltodeztrin empirically as a 
combination of an amylon group with two amylin groups, and write 

its formula as | j^i^^^i^^^^ but we think that the new light we 

have now been able to throw on the subject justifies us in adopting 
another method of statement which is more in accord with our ordinary 
views of the structure of the more simply constituted carbohydrates. 
This formula in its abbreviated form is 


Acid Hydrolysis of Starch Derivaiives, 

An accurate determination of the amount of maltose or of glucose 
which a starch derivative will yield on complete hydrolysis with 
diastase and acid respectively is of great importance in determining 
the nature and constitution of the derivative. We have so frequently 
had occasion to describe minutely the method of procedure in the 
case of diastase-hydrolysis, and the processes for estimating the 
products, that it is quite unnecessary to enter again into details on 
this point. In the course of our more recent investigations, however, 
it became necessary to make frequent use of acid hydrolysis with 
estimation of the dextA^ose formed, and we had consequently to determine 
the best conditions under which such experiments could be carried 
out, and the magnitude of the experimental errors. 

An estimation of dextrose when unmixed with any other optically 
active or reducing substance is now capable of being conducted with 
a high degree of accuracy. For the actual methods^of estimation, we 
must refer to two of our recent papers in this Journal (Trans., 1897, 
71, 72 and 275). 

If we have a substance, such as a starch derivative, which is com- 
pletely transformed into dextrose by the acid treatment, we have three 
different and independent methods of determining the dextrose formed : 
(1) by observing the increase in the specific gravity of the solution 
after hydrolysis and making up to the original volume ; (2) by the 
alteration in the optical properties of the liquid ; and (3) by the 
increase in cupric-reducing power. Theoretically, these three methods 
should give identical results, practically it was found that there are 
certain small difEerences which had to be determined and corrected for. 

One of the difficulties met with at the outset is that attendant on the 


production of more or less of r&version produotSy which invariably tend 
towards a loss of reducing power, and to a smaller extent of optical 
activity. If ^these reversion products do not exceed about 5 per 
cent, in amount they have very little influence in altering the 
specific gravity, since their solution density is very nearly that of 

Our first endeavour was to find a means of hydrolysis which would 
ensure complete conversion of the polysaccharides into dextrose with 
the minimum production of reversion products, and then to determine 
as far as possible the corrections fiecessary for these under certain 
standard conditions of hydrolysis as regards strength of acid, 
temperature of conversion, and time. After a considerable number 
of trials, with careful observations on the course of the reaction, 
we finally settled on the following standard conditions. 

A solution of the carbohydrate under experiment, of a concentration 
of from 8 to 10 per cent., is hydrolysed with 2*6 grams per 100 c.c. of 
crystallised oosalie acid. The solution is contained in a pressure flask 
immersed in a boiling water bath, and the digestion is carried on at 
100° for 14 hours. 

In determining the sp. gr. of the hydrolysed solution, this must, of 
course, be corrected for the amount of oxalic acid present, but it is 
generally better to first remove this from the hydrolysed solution 
with calcium carbonate, evaporating somewhat, and making up to the 
original volume.* 

When a definite amount of pure dextrose is treated in this manner, 
and the greatest precautions are taken to ensure accuracy in the cupric 
reduction experiments and in the weights and volume of the solutions, 
the dextrose determined by cupric reduction after digestion with the 
acid for 14 hours falls short of the amount taken by about 5 per cent. 
In one such experiment, where 9*309 grams of dextrose were taken, 
8*923 grams were estimated in the hydrolysed liquid, a loss of 4*2 per 
cent. In another experiment, 8*550 grams of dextrose before acid 
treatment were estimated as 8*187 grams afterwards, a loss of 4*3 
per cent. 

This apparent loss is due to the formation of reversion products 
which have a lower cupric reduction than dextrose. Since we know that 
these products have a higher optical activity^than dextrose, we should 
naturally be prepared to find a slight increase in specific rotation. 
This was, in fact, the case. The specific rotation of the dextrose in the 
first case increased from [a][> 52*6° to [aj^ 53*2°, and in the second 
from [a]D 52*7 to [a]i> 53-0. 

* In snoh experiments, the yolnmes are always deduced from the weighings and 
density. ^ j 

Digitized by VjOOQIC 


















On the other hand, there was no apparent change in the sp. gr. of 
the solutions before and after heating with the acid. 

With these facts in view, it is possible to apply certain corrections 
to the results obtained when carbohydrates are hydrolysed under the 
standard conditions we have described. It must be understood that 
these corrections have been applied to the results which follow. 

Hydrdyais of MdUost (anhydrous). 

Dextrose produced from 100 parts of maltose. 

1 — ^By specific gravity 104-8. 

2 — ^By cupric reduction 103*71. 

3 — ^The theoretical yield of dextrose from 

100 parts of CigHj^On is 105*26. 

Hydrolysis qf SohMe Starchy ^^e^io^s* 

1 — By specific gravity 
2 — ^By optical activity 
3 — By cupric reduction 

The soluble starch used in these experiments was submitted to a 
preliminary drying at 80°, and accurately weighed quantities of this 
were used for hydrolysis, the amount of water left in the preparation 
being determined by drying over phosphoric anhydride in the vacuum 
apparatus at 102 — 106°. The volumes of the solutions were deduced 
from the weights and specific gravity. 

In experiments I and II, the determinations were made on the 
hydrolysed solutions still containing the oxalic acid. Expt. lY 
gives the results of a further examination of II, after removal of the 
oxalic acid with calcium carbonate. 

The theoretical yield of dextrose from 100 parts of nQ^^fi^ is 

EydrolysiB of MaUodeaOrin, Q^^HaoO^ . 

The results of this experiment have been already given in the body 
of the paper, but are here reproduced for purposes of comparison. 
Dextrose yielded by 100 parts of maltodextrin. 

1— By optical activity 108-7 

2 — By cupric reduction 106*0 

The theoretical yield is 109 of dextrose. 

Digitized by VjOOQIC 


Hyd/rolysia of MaUobionic Acid, 

The test experiments on the hydrolysis of maltose and soluble 
starch show that the method gives results of a considerable degree of 
accuracy when certain corrections are applied which are constant 
under the standard conditions of experiment. It was felt desirable, 
however, to test the method also on an acid derivative of a poly- 
saccharide which has some sort of relation to the maltodextrinic acids 
we have described. For this purpose, we prepared maltobionic acid by 
the oxidation of maltose by bromine, as described by Fischer and 
Meyer {loc. cit,). Having prepared the lead salt, this was purified as 
far as possible by frequent precipitation from its aqueous solution by 
alcohol, then decomposed by hydrogen sulphide, and the free acid 
converted into the calcium salt. The calcium maltobionate was then 
purified still further by frequently precipitating with alcohol until its 
composition was constant. 

Calcium in pure salt, 5*4 per cent., against 5*3 required by 

On hydrolysing a definite weight of the dry calcium salt with 
oxalic acid, under the standard conditions, we obtained in one experi- 
ment a yield of 41 '6 per cent, of glucose, and in a second experiment 
42-6 per cent. The theoretical yield of glucose from {0i^^\0^^jd9k 
on complete hydrolysis is 47*7 per cent. 

Dayt-Fabaday Resbabch Laboratory, 
Royal Institution. 

XXXL — Attempts to Prepare Pure Starch Derivatives 
Through their Nitrates, 

By Horace T. Brown, LL.D., F.R.S., and J. H. Millar. 

The comparative ease with which certain nitrated derivatives of 
carbohydrates can be obtained, and the ready manner in which the 
NOg-groups can be subsequently removed by treatment with ammonium 
sulphide, induced us, some time ago, to turn our attention to a 
method which we hoped could be made applicable to the separation 
of the dextrins or maltodextrins which occur in restricted starch 
transformations by diastase. These substances differ so little in 
their relative solubility in alcohol of various strengths that the 
ordinary modes of separation are extremely tedious and difficult. 
The two important points to consider were the degree of ease with 
which the nitrates could be fractionated by treatment with various 

Digitized by VjOOQIC 


solvents, and how far one could be sure that the recovered product 
obtained by the ammonium sulphide method represented the original 
substance before nitrification. 

Although our hopes have only been partially realised^ the work has 
not been without practical value, since by its aid, as will be seen in 
the following paper, we have obtained some indirect evidence as to 
the nature of the ' stable ' dextrin of starch transformations. 
Apart from this comparatively small result, we believe that the work 
should be put on record as a guide to others who may be induced, 
like ourselves, to give a considerable amount of time to a subject 
which at the outset looks a promising one. 

It is only by applying the method of nitration and recovery to 
starch derivatives of well-defined properties that we can hope to test 
its value. We have consequently applied it to (1) soluble starch, 
(2) the * stable * dextrin after purification with alcohol, (3) amylo- 
dextrin, (4) maltodextrin, and (5) maltose. 

Miration a/nd Recovery qf Soluble Starch, 

The soluble starch was carefully prepared from potato starch by 
lintner's acid process. Twelve grams were nitrated at 0^ with 70 c.c. 
of strong nitric acid, and the nitrate was precipitated in the first 
place with 70 c.c. of strong sulphuric acid, and afterwards with water. 
It was purified by dissolving it in ether and precipitating with 
chloroform. Analysis of the product by Lunge's method showed 

1. — 9*3 per cent, of nitrogen. 2. — 9'1 per cent, of nitrogen. 

The trinitrate of the empirical formula 022^17^7(^^8)3 i^^^ures 
91 per cent, of nitrogen. 

Soluble starch was recovered from this nitrate by treating it at 
70^ — 80% with water containing a little ammonium sulphide, and pass- 
ing through the liquid a stream of hydrogen sulphide ; under these 
conditions, the nitrate gradually dissolves with separation of sulphur, 
which, however, again goes into solution in an excess of hydrogen 
sulphide. The clear solution was then heated to the boiling point 
and the precipitated sulphur filtered off, using a hot funnel. When 
the concentrated filtrate was treated with an equal volume of 90 
per cent, alcohol, a thick, slightly discoloured, flocculent precipitate 
was thrown down ; this was re-dissolved in water and again precipi- 
tated with alcohol, the process being repeated several times until the 
product was obtained as a perfectly white, amorphous powder. The 
recovered starch had all the physical properties of ordinary soluble 
starch, giving the same pure deep blue colour with iodine. Its 
specific rotatory power was [ a ]i> 201*2°. 

Digitized by VjOOQIC 


When a portion of this recovered starch was hydrolysed with malt 
extract at 56^, it behaved exactly as ordinary soluble starch does, the 
mixed products of transformation having a specific rotatory power of 
[a]]> 151*8 and a cupric-reducing power of B 76*4. 

The recovered product is, in fact, ordinary soluble starch in a state 
of great purity, and absolutely free from ash. 

This process for the preparation of soluble starch can be recom- 
mended when a very pure specimen is required. 

NiiraUon a/nd Attempted Recovery of the StaNe Dextrin, 

The dextrin from which these preparations were made was obtained, 
as described in the next paper (Trans., 75, p. 317), by repeated alcoholic 
treatment of starch products hydrolysed to the lowest possible point 
by diastase. It is the stable dextrin of the so-called ''No. 8 
equation," frequently referred to in our previous papers. The pro- 
perties of this dextrin will be fully described in a following 
communication. It was finally obtained with the following optical 
and reducing values : [a]D 195*0 ; R 5*5. 

On treatment with nitric acid, exactly as described in the case of 
soluble starch, this dextrin yielded a nitrate in the form of a per- 
fectly white powder, insoluble in water, but soluble in alcohol and 
ether or in a mixture of the two solvents. From such solutions, it is 
completely precipitable by chloroform. Fractionation of the product 
showed that it must be a mixture, since it was divisible into two 
portions containing 10*1 and 9*36 percent, of nitrogen respectively. 

Assuming the dextrin to have a composition corresponding to 
nC^HjQOg, its di-, tri-, and tetra-nitrates would require the following 
percentages of nitrogen. 

Dinitrate 6*7 per oent. N. 

Trinitrate 9*1 „ ,, 

Tetranitrate 11*1 „ „ 

It now remained to be seen if the original dextrin could be 
recovered from the mixed nitrates by applying the ammonium 
sulphide process, which had been so successful in the case of soluble 

Owing to the recovered dextrin requiring stronger alcohol than 
soluble starch to throw it out of solution, it is much more difficult to 
completely free it from ammonia salts and fixed ash. These were 
finally removed by dialysis, and the recovered ' dextrin' was precipi- 
tated several times with alcohol until its optical properties were 

The optical and reducing values for the recovered product are here 

„. „„ , ^oogfe 


given for two distinct preparations, the values for the original 
dextrin being also given for comparison. 



Original dextrin 



* Dextrin ' recovered from nitrates. 

Prep. (1) 


Prep. (2) 



Although at first sight it may appear that a non-reducing dextrin 
of high optical activity has been recovered, such is really not the 
case. If the reducing portion of the original dextrin with an B of 
5*5 had been merely separated in the process of nitrification and 
fractionation, we should expect the specific rotatory power of the 
recovered ' dextrin ' to be correspondingly increased, instead of being 
materially decreased, as it in fact is. 

The distinct acid character of the recovered ' dextrin ' is a proof 
that the original dextrin had been oxidised by the treatment with 
nitric acid, and that we have been really dealing with the nitrate of 
a polysaccharide acid, and not with that of the original dextrin, which 
is without acid characters. 

This is a matter which will be found more fully dealt with in the 
next paper ; it is sufficient for our present purpose to know that the 
nitration and recovery method is not applicable to the purification of 
the stable dextrin. 

NUrcUion and AiUmpted Recovery of AmylodextHn, 

The name amylodextrin was applied by W. Nageli in 1874 to a 
substance which he obtained by the long-continued action of dilute 
mineral acids on ungelatinised starch in the cold. Unfortunately, of 
late years, a considerable amount of confusion has been introduced 
into the nomenclature of starch products by an application of the 
same name by C. J. Lintner and others to another substance which is 
identical with the 'soluble starch ' of O'Sullivan, Musculus, and 

In 1889, one of us and Morris (Trans., 1889, 65, 449) described the 
results of an investigation of Nageli's amylodextrin, and it was 
shown to be a well-defined substance belonging to the class of malto- 
dextrins or amyloins. It separates from its solutions in the form of 
sphflsro-erystals, and has optical and reducing properties correspond- 
ing to [a]y8-8« 206-1 and icsse 9*08 (= [ajo 192-0°, R 14-9) 

It was regarded as constituted of six amylin groups combined 
with one amylon gronp, and consequently having the empirical formula 

I 0„H^O„ Digitized by Google 


In the light of our recent work on maltodextrin, described in 
the previous paper, the constitution of amylodeztrin will doubtless 
be expressed in the abbreviated form by 

O^^l 2^20^9 

A part of the actual preparation of amylodeztrin described in the 
1889 paper was submitted to nitration under the usual conditions. 

4*800 grams yielded 7'78 grams of nitrate (162 percent.), which came 
down on successive treatment with sulphuric acid and water as a white, 
amorphous powder, very soluble in ether, acetone, and ethy lie acetate, 
but insoluble in chloroform and light petroleum. 

Three fractional precipitations were made with chloroform from its 
solution in etbylic acetate. These gave the following amounts of 

Fraction (1) 10-81 per cent. N. 

(2) 10-66 „ „ 

(3) 11-50 „ „ 

Supposing it to be a true nitrated derivative of the unaltered 
amylodeztrin, this corresponds very nearly with the 28-nitrate, which 
requires 11*0 per cent, of nitrogen. 

The nitrate was treated in the usual manner with ammonium 
sulphide, and the recovered amylodeztrin precipitated with alcohol ; 
it was obtained perfectly free from ash, and in the form of minute 
sphserocrystals. The optical and reducing properties of the original 
and recovered products are here given. 

[o]d R. 

Original amylodeztrin 192*0 14*9 

Amylodeztrin recovered from nitrate 181*0 8*1 

It is evident that the substance has been much altered during 
the process, and that the method of recovery is not applicable to 

Ntiration and Attempted Recovery qf McUtodeoetrin, 

4*686 grams of carefully dried maltodeztrin, on nitration, yielded 
7-094 grams of the nitrate (151 per cent.). This is very soluble in 
85 per cent, alcohol, and also in ether and ethylic acetate. From its 


ethereal solution, it is thrown down by light petroleum in the form of 
oily dropS; but is only partially separable by chloroform. Chloroform 
does not precipitate it from its solution in ethylic acetate, and light 
petroleum only partially. 

The nitrate was fractionated by partial precipitation with light 
petroleum, evapyration of the mother liquors, re-solution in ethylic 
acetate, and further precipitation with light petroleum. Two 
fractions prepared in this manner gave the following amounts of 
nitrogen by Lunge's method. 

Fraction (1) 7*80 per cent. N. 
(2) 10-34 „ „ 

The mixed nitrates, on treatment with ammonium sulphide, gave a 
substance which had the appearance of maltodeztrin, but was of an 
acid nature and had been reduced in specific rotatory and reducing 
powers to 

Wd 185-1 
R 26-3. 

Alterations have, therefore, occurred during nitration, and the 
recovered product is not identical with the original. 

Ni^^aiion cmd Attempted Recovery of Maltose. 

The nitrate was readily obtained as a white, amorphous solid, which, 
after well washing with water, was re-dissolved several times in 
ethylic acetate and precipitated by chloroform, which brings it down 
as a syrup. This was separated, and after partial drying, the product 
was powdered and completely dried until the weight was constant. 
Lunge's method showed 13*9 per cent, of nitrogen in the purified 
substance. The hezanitrate, O^JELiQ^ii^^sief requires 13*72 per cent, 

This nitrate did not yield pure maltose on treatment with ammonium 
sulphide, the specific rotatory power of the solution being very much 
higher than it should have been if the cupric reduction were taken as 
a measure of the maltose present. 

Nevertheless, the presence of maltose in the mixture was ascertained 
by the formation of maltosazone. 

General Remarke. 

It will be observed that the regeneration of the original starch 
derivative in anything like a pure state from its nitrate was only 
found possible in one case, that of soluble starch. In all those 

Digitized by VjOOQIC 


instances in which the starch derivative under experiment was cupric- 
reducing — that is, contained the open carbonyl group — ^the properties 
of the regenerated product indicated that more or less oxidation had 
taken place, and that the ethereal nitrate was an indeterminate 
mixture of a nitrate of the original carbohydrate and of a nitrated 
carboxylic acid or acids. This line of work, commenced in the early 
part of 1897, was consequently abandoned, since it did not promise 
to lead to the object we had in view, the separation and identifica- 
tion of the intermediate products of starch-hydrolysis. 

Since that time an important paper on the nitration of carbo- 
hydrates has been published by Will and Lenze {Ber., 1898, 30, 68) 
in which they give an account of their examination of most of the 
principal carbohydrates from this point of view. They found that, 
with the exception of the lactoses, most of them are readily converted 
into nitrates, and in many cases these were obtained crystalline, and 
with such well characterised properties, that the process may be used 
for identification of uncertain carbohydrates and their recognition in 
mixtures. An attempt to regenerate the original substance appears 
to have been made only in one instance, that of Fischer's a-methyl- 
glucoside, whose tetranitrate yielded the parent substance when an 
alcohob'c solution was treated with ammonia and hydrogen sulphide. 

Of the substances described in our present paper, the authors 
examined only starch and maltose. The former gave an amorphous 
hexanitrate, whilst from the nitrated products of the latter a crystalline 
substance was obtained whose nitrogen content corresponded to 
maltose octonitrate. 

Had we been successful in obtaining definite nitrates of the inter* 
mediate starch derivatives, it was part of our plan to determine their 
molecular weight by the freezing point method. We have, in fact, 
made these determinations, but from what has been said it is evident 
that they are without value, and we consequently have not given 
them here. 

We may, however, refer to the examination of the derivative of 
soluble starch from this point of view, since everything pointed to the 
derivative in this case being a perfectly definite trinitrate. 

The determination was made in acetic acid, whose constant of 
molecular depression had been determined by means of naphthalene. 

The result gave a value for Ma>987. The molecular weight of 
2[Ci2HiA(NO«)3] is 918. 

From all we know about the constitution of soluble starch from 
other sources, it is quite certain that the value of n in the formula 
n(Oj2H2oOio) must be far greater than 2, so we must conclude that 
the freezing point method applied to such derivatives of the colloidal 

Digitized by VjOOQIC 


carbohydrates as we are considering is quite inapplicable to the 
determination of molecular weight; and the same probably applies 
to the colloidal carbohydrates themselves. 

The Davy-Fabaday Reseakch Laboratoky, 
Royal Institution. 

XXXII. — The Stable Dextrin of Starch Transformations, 
and its Relation to the Maltodextrins and Soluble 

By Ho&ACB T. Bbown, LL.D., F.R.S., and J. H. Millae. 


I. Intiodnction 315 

II. Separation and Purification of the Stable 

Dextrin 317 

HI. Is the Stable Dextrin a Reducing Substance 9 . 322 
lY . Oxidation of the Stable Dextrin and Formation 

of Dextrinic Acid 325 

y. Nitration and Subsequent Recovery of Dextrinic 

Acid 326 

YI. Analysis of Calcium Salt of Dextrinic Acid . 328 
yn. Hydrolysis of Dextrinic Acid with Oxalic Acid 

and Diastase 328 

ym. Hydrolysis of the Stable Dextrin with Oxalic 

Acid and Diastase 330 

IX. Generalisations and Conclusions . . 331 

I. Introduction. 

It was shown by one of us and Heron in 1879 (Trans., 1879; 36, 596, 
that when gelatinised starch, or soluble starch, is hydrolysed with a 
little cold-water malt-extract at temperatures below 60°, the reaction 
proceeds with great rapidity until the mixed products of transforma- 
tion show a cupric reduction corresponding to about 80 per cent, of 
maltose, and a specific rotation very close on [a}i> 150°.* With an 
active diastase, this point may be reached in less than 5 minutes, and 
further change is comparatively very slow indeed. 

* At oidiiMty tempentaies, the tpecific rotatorylpower may be temporarily lower 
than thia owing to the maltoee being liberated in the birotatory state. (See Trans. , 
1895, 67, 809.) 

Digitized by VjOOQIC 


This fact has been referred to many times in our papers, and was 
illustrated in the 1879 paper by a series of time carves, and again in 
the same manner in a paper on amylodextrin published in 1889 
(Trans., 65, 457). 

These values of [ajo 150 and B 80 mark, in fact, a definite resting 
point in the reaction, beyond which it is difficult to push it, unless a 
considerable amount of time is allowed. 

That the cessation of action is due neither to the weakening of the 
enzyme nor to the action being a reversible one,* may be readily shown 
in a variety of ways. For instance, (1) the addition of more enzyme has 
no appreciable effect in hastening the very slow subsequent change 
after the resting point has been reached ; (2) if more gelatinised 
starch or soluble starch is added to the solution, it is speedily brought 
to the same condition of optical activity and reducing power; (3) 
when a complete or partial separation of the maltose is effected by 
suitable means, as by fermentation or fractionation with alcohol, the 
residual dextrin shows the same resistance to further hydrolysis with 
diastase as^o the mixed products which have attained the resting 

In the 1879 paper {loc. cit), it was also shown that, if we assume 
the whole of the reducing substance to be maltose, the resting point 
of the reaction corresponding to [aJD 150° and It 80 agrees very 
closely with the view that the starch has been hydrolysed according 
to the following empirical equation, 

lOCijHjoOio + SKfi = 8C12H22O11 + 2C12H20O10 
Starch. Maltose. Dextrin. 

which requires 80*8 per cent, of maltose in the products of trans- 

This was the so-called ' No. 8 equation,' which at that time was 
supposed to represent only one of several other halting points in the 
hydrolysis. We are now of opinion that this well-marked resting 
stage, corresponding to [ajo 150, R 80, is the only one which admits, 
with certainty, of being expressed by any definite equation. 

From the solution of the starch products which have been allowed 
to reach the above-mentioned point, crystallisable maltose can readily 
be prepared by extraction with alcohol of 80 per cent., or the maltose 
can be fermented out by yeast. By either of these methods of treat- 
ment, by far the greater part of the reducing substance may be 
removed, and in this respect these complete starch transformations 
stand in strong contrast with those in which the resting stage has not 

* By this we do not mean to imply a denial that the action may not be to some 
extent a reveraible one, but only that this particalar effect cannot be due to 

Digitized by VjOOQIC 


been fully attained, since, in all incomplete conversionB, the reducing 
constituent is not present wholly as crystallisable, fermentable maltose, 
but, in part, at any rate, as maltodeztrins. Complete and incomplete 
conversions also differ from one another in their behaviour towards 
phenylhydrazine. If the transformation products have reached the 
resting stage corresponding to [aj^ 150, R 80, the yield and purity 
of maltosazone is much greater than is obtained from a proportional 
amount of an incomplete transformation which might be assumed, 
from its reducing power only, to contain an equal amount of maltose. 
Indeed, all the available facts bear out the supposition that in a 
complete conversion which has attained the resting stage, the greater 
part, if not the whole, of the reducing power is due to free, crystallis- 
able, and readily fermentable maltose. 

Such a solution, besides maltose, contains an achroodeztrin, 
which is precipitable by alcohol of 80 — 85 per cent. This dextrin, 
amounting to about 20 per cent, of the starch originally taken, is 
attacked only with difficulty by diastase, and differs in this respect 
from all the intermediate dextrinous products, amongst which is 
the maltodeztrin which has been the subject of a previous paper. 

We have submitted this stable dextrin to a careful re-examination 
during the past two years, and the present paper deals with our most 
recent conclusions with regard to its nature, properties, and con- 

The difficulties of obtaining the stable dextrin in a state of 
purity are considerable, and even greater than they are in the case 
of maltodextrin. This arises from the fact that the process of 
treatment with alcohol necessarily tends to concentrate in the final 
insoluble product all impurities pre-existent in the starch and the 
transforming agent, which happen to be insoluble in the lowest 
strength of alcohol used in the separation. 

If fermentation has also been employed, there is the additional 
risk of introducing small amounts of nitrogenous impurities, and 
also some of the non-volatile products of fermentation, although the 
latter are, for the most part, removable by the subsequent treatment 
with alcohol. 

The ^actual amount of such impurities, which may amount to 
2 — 3 per cent., can be determined with a close approximation to 
accuracy by submitting the separated dextrin to the standard 
methods of acid hydrolysis described in the'previous paper (p. 305). 

II. Separation oflnd Parificatum of the StaUe Dextrin. 

In these experiments, we always commenced with from 2000 to 
3000 grams of well-wasl)ed potato starch, which was brought into 

VOL. I^XV. Digitized by (SoOglC 


solution with from 16 to 20 litres of water by successive gelatin!- 
sations and liquefactions, in the manner described in our previous 
paper on maltodextrin (this vol., p. 288). We thus obtained the 
starch solids, for the final conversion, at a concentration of from 
12 to 15 per cent. The preliminary conversions at 65 — 70® were 
allowed to go sufficiently far to prevent any of the higher starch 
transformation products coming out of solution when the tempera- 
ture was reduced to 15 — 20^ The completion of the conversion 
down to the ^ resting point/ corresponding to [a]D 150®, R 80, was 
then carried out at the ordinary temperature, with the aid of cold- 
water malt-extract, made from a very diastasic air-dried malt. This 
conversion sometimes lasted for two days * or more, the object being 
to select the best conditions for producing a low conversion, with the 
employment of the minimum amount of malt-extract. The importance 
of this condition is rendered evident by what we have already said 
about the concentration of impurities on the dextrin during the sub- 
sequent processes of its extraction. 

The actual amount of ' normal ' t malt extract required was from 
3 to 6 c.c. of the extract per 100 c.c. of a solution containing from 
9 to 15 grams of starch products. This is equal to from 0*3 to 0*4 
c.c. of extract per gram of starch converted, and represents about the 
minimum amount which can be used for these complete conversions. 
The actual amount of solids introduced with the malt -extract repre- 
sents about 3*5 per cent, of the starch used, but by far the greater 
part of these solids are separable by alcohol of from 75 — 90 per cent., 
and are removed in the processes of purification to which the dextrin 
is submitted. 

After the conversion had been run down to the desired point, the 
solution, filtered from a certain amount of ' starch-cellulose,' which 
always separates during starch transformations made in the cold, 
was evaporated to a thick syrup. The process of separation was 
then commenced; this consisted of fractionation with alcohol of 
various degrees of strength with occasional recourse to fermenta- 
tion. The treatment with alcohol was varied in every conceivable 

It would be tedious and unprofitable to give anything but a mere 
sketch of the process, since, at this stage, it did not differ materially 
from similar methods we have frequently described before, except as 
regards the number of operations and the length of time required, 
often amounting to severed months for each experiment. 

• The solution was occasionally sterilised, with the further addition of malt 

t ' Normal * malt extract is obtained by digesting finely divided malt (air dried) 
with 2*5 times its weight of cold water, and filtering after 6 hours. 

,.gitized by 



We will briefly describe the coarse of one experiment, which is 
typical of all. 

A conversion of 2400 grams of starch which had been run down 
to the ' resting stage ' and evaporated to a syrup was treated in the 
first place with 7 litres of hot alcohol of 90 per cent.,* the strength 
of the alcohol in the mixture being 75 — 80 per cent. 

About 200 grams of dextrin separated. The alcoholic solution, 
poured off when cold, was distilled, and the syrupy residue poured 
into 6 litres of boiling spirit of 96 '5 per cent., and digested for some 
time. The second dextrinous residue thus obtained was further 
digested with 86 per cent, spirit, the insoluble portion being added 
to the original residue of 200 grams. The alcoholic solutions on 
distillation yielded a syrup from which maltose readily crystallised. 

The combined dextrinous portions, amounting to 660 grams, now 
contained only 42 per cent, of maltose, as against 80 per cent, in the 
original starch transformation products. This crude dextrin was 
dissolved in 6 litres of water, and the solution, having a sp. gr. of 
1042*6, was set to ferment with 16 grams of washed and pressed yeast. 
During the progress of the fermentation, which lasted for 10 days, 
an examination of the solution was made from time to time, and it 
was shown that, whilst the specific rotatory power of the unfermented 
matter steadily increased, the optical properties of the portions 
fermented at any stage corresponded with those of maltose. 

The fermented solution, which now gave only a trace of insoluble 
osazone on treatment with phenylhydrazine, was evaporated to remove 
alcohol. The specific rotatory power of the residue was [a ]j, 182*5°. 

In order to remove any trace of maltodextrin, the solution, made up 
to sp. gr. 1043, was now once more treated for 2 hours at 50° with 
40 C.C. of a very active malt extract. It was then once more heated 
to the boiling point, cooled, and again set to ferment with 11 grams 
of washed yeast for 5 days, during which time 14*04 grams of ferment- 
able substance disappeared, which represents about 3*0 per cent, of 
the 540 grams of the crude dextrin. The optical properties of this 
fermentable substance were again found, by the usual process of 
fractional fermentation, to correspond with those of maltose. 

Two precipitations were then made with 85 per cent, alcohol, and 
the dextrin obtained was submitted to nine successive extractions 
with boiling alcohol, each extraction lasting for a day or more, the 
mixture being frequently agitated. The first six of these extractions 
were made with alcohol of 86 per cent., and the other two with 
alcohol of 80 per cent. The weights and specific rotatory powers of 
the matter extracted, and the specific rotatory power of the residual 
dextrin after each extraction, were determined. 

* Throughout the paper, the percentages of alcohol are expressed by volume. 


The results of the nine successive treatments at this stage are here 

The amount of crude dextrin at the commencement was 492 grams. 

Table I. 

No. of fraction. 

Extracted with alcohol. 

Residual dextrin. 





00 ^6 









22 12 


17 11 


127-8 * 



167 1 












* These extractions contained the non-Tolatile prodncts of fermentation and the 
sugars of the malt-extract. 

It was evident that the method of separation adopted up to this 
point had practically reached the experimental limits, so a different 
plan of treatment was now adopted. 

The dextrinous residue, after the ninth treatment, amounting to 
383 grams of substance with an [ajo of 186*8°, was dissolved in 
1000 c.c. of water, and cdd alcohol of 85*5 per cent, was gradually 
added, until precipitation just commenced. The precipitate containing 
albuminous matter was rejected. Alcohol of 85-5 percent, was again 
added with constant agitation, until about one-third of the dextrin 
was thrown down. This is fraction A, and it was separately treated 
by dissolving it in its own weight of water, and precipitating in two 
fractions, A^ and Ag, by the further addition of 85 per cent, spirit. 

These fractions had the following properties. 

Ai [a]o 193-1 R 6-6 
Ag [a],> 193-0 — 

The clear solution from A was again precipitated in the cold with 
alcohol. On heating, the precipitate redissolved, but was again thrown 
out on cooling — this is fraction B. After the removal of the precipi- 
tate, the liquid was increased in alcoholic strength, in the first place 
to 85 per cent., and in the second to 95 per cent., the fractions thus 
obtained being C and D. The final mother liquor from D was mixed 
with that from A^, and the 55 grams constituting fraction E were 

Digitized by VjOOQIC 


freed from alcohol and separated into two parts, E^ and E^, with 
95 per cent, alcohol. These latter fractions had the following specific 
rotatory power. 

Fraction B [a]i> 192-1° Fraction E [a]i> 184-9° 

C [a]D 190-8 Ej [a]i, 186-6 

D [aJD 189-0 Ej [a],> 179-7 

It is clear from the above that the dextrin which could not be 
further separated by the nine treatments with alcohol described in 
Table I has given way to the modifications in the alcohol treatment 
just described, products being obtained amongst the least soluble 
portions with specific rotatory powers as high as [aj^ 193-0° 

The above fractions A^, Aj, B, d and D, with rotatory powers of 
from [a]]> 189*0 to [a]i> 193*1 were now united and subjected to still 
further treatment. 

After being evaporated to a thick syrup, they were gradually treated 
with 2 litres of hot 86 per cent, alcohol ; after digestion for 6 hours, 
the alcoholic solution, when cold, was poured off, and the properties 
of the residue, and those of the small quantity of extracted matter, 
were examined. These were as follows. 

Dextrinous residue [ajn 192*7° 

Portion soluble in alcohol... [a]^ 183*6 

This process was repeated twice more, and the dextrin, which had now 
been reduced to about 200 grams, was obtained with a specific rotatory 
power of [a]D 192-0°. 

We now once more resorted to fractionation. The 200 grams of 
dextrin were dissolved in 500 c.c. of water, and 500 c.c. of alcohol added in 
the cold ; the addition of 100 c.c. more of 96 per cent, alcohol brought 
down fraction F ; this, when redissolved in 200 c.c. of water and again 
treated with 250 c.c. of 96 per cent, alcohol, gave a precipitate which, 
by further similar treatment, was resolved into fractions F^, F3, and 
F^, having the following optical and reducing properties. 












It must be remembered that, in the protracted and laborious treatment 
of separation, which ishere only described inavery condensed form, every 
conceivable variation of treatment with alcohol was adopted, whilst 
the process was further supplemented by occasional treatment with 
diastase to break down any maltodextrin which might remain adherent 
to the product, and also by employing fermentation to assist in the 
elimination of the maltose. ^ ^ 

Digitized by VjOOQIC 


In all similar experiments, no matter how long the alcoholic frac- 
tionation may have heen continued, we have always obtained the same 
final result. A point is ultimately reached, sometimes comparatively 
speedily, when no further change takes place in the properties of the 
separated dextrin. This is marked by the product assuming optical 
and reducing properties corresponding to 

[a]Dl95— 196« 
R 5-5— 6-9 

Such a product is, however, never quite free from ash, which amounts 
to 0*3 to 0*5 per cent, of the dry weight, and it also contains a trace 
of nitrogen corresponding to about 0*2 per cent, of albuminoids. 
Hence the above values of [ a ]d and R are slightly lower than those 
corresponding to the pure carbohydrate ; the error introduced from 
this cause does not, however, exceed more than 2° in the specific 
rotatory power, and will have no appreciable effect on the value of R. 

Prepared in this manner, this dextrin is obtained as a white, 
amorphous, gummy substance, soluble in water to any extent, and 
completely separable from its aqueous solutions in the cold by alcohol 
of 80 per cent., which brings it down as a waxy solid with a silky 

Dried at 100^ in a vacuum over phosphoric anhydride, it has a 
divisor of 3*995 for a solution density of 1054*49. 

One of the chief characteristics of this dextrin, prepared from starch 
conversions which have attained the * resting stage,' is the relative 
difficulty with which it is attacked by diastase, as compared with the 
small power of resistance exhibited by the intermediate products. 
The mode in which it breaks down under the agency of dilute acids, 
and the long continued action of diastase, will be described later on. 

III. 1$ the Stable Dextrin a Reducing Substance f 

We have already stated that the final product obtained by pro- 
tracted treatment with alcohol, assisted by fermentation, always 
possesses a reducing power of about R 5*5, and we have now to con-' 
sider whether this is inherent to the dextrin itself, or whether it may 
not, after all, be due to an admixture of from 10 to 15 per cent, of 
compounds of the maltodextrin class which have persistently adhered 
to it through all the long processes of treatment to which it has 
been subjected. The reducing power certainly cannot be* due to 
maltose itself, since the purified dextrin gives no trace of maltosazone 
when treated with phenylhydraaine, neither does it contain any 
substance directly fermentable with yeast. 

If the stable dextrin in a state of purity is a non-reducing substance, 


it is possible to deduce its optical properties from an analysis of the 
mixed products of a starch transformation which has been carried 
down to the ' resting stage.' With the improved methods of analysis 
which we have described in a previous paper (Trans., 1897, 71, 72), it 
is possible to determine this value with considerable accuracy, for 
under these conditions complete corrections can be made for the 
transforming agent (malt-extract) employed, and for the small amount 
of ash existing in the original starch. After fractionation, these cor- 
rections cannot, of course, be applied with the same degree of accuracy. 

We have made two such experiments for the determination of the 
specific rotatory power of the stable dextrin, on the assumption that 
it is non-reducing. 

It is necessary, in the first place, to determine very exactly the 
total weight of the starch products per 100 c.c. of the liquid after 
complete transformation. This is effected by correcting the specific 
gravity for that due to the malt-extract, and then employing the 
particular < divisor ' corresponding to the phase of transformation 
and the concentration of the solution. 

The ' divisor curve ' for a transformation corresponding to [ a ]d 150, 
R 80 has been accurately determined for concentrations of from 
2*4 to 22*3 per cent, of solids, and will be found at p. 80 of our 
previous paper on << Experimental Methods'' (Trans., 1897, 71, 72). 
In such experiments as we are describing, where the highest degree of 
accuracy is necessary, a further small correction was necessary for 
the 0*2 per cent, of ash which the original starch contained.* 

The amount of apparent maltose per 100 c.c. was then determined from 
the cupric reduction, due correction being, of coarse, made in all cases 
for the reduction due to malt-extract. 

The amount of rotation corresponding to the maltose was then 
calculated from the known optical constants of the maltose at the 
particular concentration, and the value thus obtained was deducted 
from the total (corrected) optical activity of the original solution. 
The remainder represented the rotation due to the dextrin, the 
amount of which was also known by subtracting the known amount 
of maltose from the total (corrected) starch solids per 100 c.c. 

We thus have all the elements necessary for the calculation of the 
specific rotatory power of the dextrin. 

To ensure the greatest possible accuracy, the volumes of the liquid 
were deduced from weighings and densities, and the cupric reductions, 
several times repeated, were also made on weighed amounts of the 

* This correction only aflects the divisor in the third decimal place. For instance, 
the divisor from the onrve being 3*958 for sp. gr. 1038*61, becomes, when corrected 
for 0'2 per cent, of ash, 8*964. ^ j 

Digitized by VjOOQIC 


In the first experiment, the mixed transformation products corres- 
sponded to [a]© 160-3, R 79-6, and in the second to [a]© 150-4, 
R 79-0. 

The estimated specific rotatory power of the stable dextrin, 
(tsswmmg U to be non-redueingy came out as follows. 


g)[:J:X-?° )»»•»'■ 

The properties of the dextrin which we obtained by continuous 
fractional treatment of the starch transformation products corres- 
ponded, as we have seen, to R 5*5 to 5*9 and [ajo 195° to 196°, with a 
possible error of + 2°. These properties are quite consistent with a 
mioshbre of 94*5 per cent, of a non-reducing dextrin withan [a]i> 200*9°, 
and 5 *5 per cent, of maltose or its equivalent of maltodextrin. Theoptical 
properties of such a mixture would correspond to [aj^ 197*4°. We 
are, however, by no means justified in concluding that the dextrin 
obtained really consists of such a mixture, and must seek further 
evidence in other directions. 

In the paper of 1885 on the ^* Non-Crystallisable Products of Starch 
Transformation,'' by one of us and Morris {loc, cit,), a process was 
described by which the residual reducing power of dextrins, previously 
separated by alcoholic fractionation, could be completely destroyed, 
and the apparently unaltered dextrin recovered in a non-reducing 
state. This method consisted in acting on the dextrin with Knapp's 
mercuric cyanide solution, and subsequently acidifying with hydro- 
chloric acid and separating the mercury with hydrogen sulphide, the 
dextrin being afterwards separated from the filtrate by precipitation 
with alcohol. Since the optical properties of the dextrins were 
somewhat increased by this process, and their behaviour under 
hydrolysis with diastase was apparently unaltered, it was believed 
that a separation had been effected between the dextrin and the 
substance to which it originally owed its reducing power. 

In 1890, Scheibier and Mittelmeier, in a valuable and highly 
suggestive paper on starch products {Ber., 1890, 2S, 3060), objected 
to this method on the ground that, if the dextrin were in itself a 
reducing substance, its aldehydic group would necessarily be oxidised, 
and that the resulting product would be an <icidy and not the original 
unaltered dextrin. Unfortunately, the value of the experimental 
portion of the paper is much lessened by the fact that Scheibier and 
Mittelmeier made use of a commercial dextrin, which from their 
description must have been a torrqfaction product of starch, at all 
times a very indeterminate substance, whose nature and relation to 
the dextrins of starch transformations has never been^worked out. 

„.gitized by Google 


Nevertheless, the criticism of the method we employed to destroy the 
reducing power is perfectly just and well founded, as will be seen by 
a perusal of our recent paper on maltodeztrin (this vol., p. 286), 
which contains the first experimental proof that a reducing starch 
product can give rise, on oxidation, to well-defined polysaccharide 
acids. It was the fact that these complex acids have such feeble 
acid properties, and resemble the parent substance so closely in 
optical properties and in their behaviour towards diastase, which 
caused us to overlook their formation in our work of ten years 

In the light of this work on maltodextrin, it appeared to us that 
the question whether the reducing power of R 5*5 possessed by the 
stable dextrin after repeated treatment is really inherent to the dex- 
trin molecule, and is not due to admixture with something else, ought 
to be settled by oxidising the dextrin with mercuric oxide and baryta 
to the point of disappearance of its reducing power, and then once 
more fractionating the product. If the reducing power is due to a 
foreign substance, it is this alone which will be oxidised, and the acid 
or acids so formed ought to be separable from the dextrin which 
will be left as a neutral substance. If, on the other hand, the 
reducing power is inherent to the dextrin molecule, we ought to 
obtain a complex ' dextrinic acid ' which will not give way to 

We have carried out this experiment several times with results 
which, we think, leave no doubt as to which supposition is correct. 

IV. OxtdcUion qf the Stable Dextrin cmd Formation of Dextrinic 


A preparation of the stable dextrin with [ajo 195*0, E 5*5 was 
oxidised with mercuric oxide and barium hydroxide under the condi- 
tions employed for the oxidation of maltodextrin (this vol., p. 293). 
It was found that 70 grams of the dextrin with an excess of mercuric 
oxide required the gradual addition of 4*27 grams of crystallised 
barium hydroxide, Ba(0H)2 + SH^O, to completely destroy its power 
of reducing Fehling's solution. 

After filtration, the solution, slightly acidified with sulphuric acid, 
was again filtered, and the product was recovered by several precipi- 
tations with alcohol of 80 per cent At this point, the oxidised dextrin 
had a specific rotatory power of [a] d 192*5^. 

It was then partially precipitated three times successively with 
alcohol of 75, 84, and 85 per cent. The properties of the several 
fractions were 

Digitized by VjOOQIC 




1. 193-2 

2 1920 

3. 1920 

Although quite free from sulphuric acid, the solutions were distinctly 
add to test paper. 

It was now attempted to split this oxidised dextrin further hy dis- 
solving it in 50 per cent, alcohol and fractionally precipitating with a 
5 per cent, solution of barium hydroxide in 50 per cent, methylic 
alcohol; the precipitated barium compounds were then separately 
treated, first with carbonic anhydride, and afterwards with dilute sul- 
phuric acid in the cold, the oxidised dextrin being recovered from 
>he filtrate by precipitation with alcohol. 

The products thus obtained had specific rotatory powers varying 
only between [a]© 193-2° and [a],, 193-6° E 0. They were quite dex- 
trinous in their physical characters, but all had a very distinct acid 
reaction, from which the parent dextrin was quite free. 

The oxidised dextrin, which we shall in future refer to as dextrinic 
add, when dried in a vacuum over phosphoric anhydride, gave a divisor 
of 4-000 for a solution density of 1039*0. It was not quite free from 
ash, so at this stage it was not found practicable to determine its 
molecular weight from its barium or calcium salt. 

On hydrolysis with oxalic acid,* 100 parts of dextrinic acid yielded 
105 '9 parts of glucose, as against 108*8 parts yielded by the original 
dextrin {vide ir\fra). 

V. Nitration amd Subseqwnt Recovery of Dextrinic Add, 

We have shown in a previous paper (this vol., p. 310) that when 
the stable dextrin is nitrated, the nitrate, on treatment with 
ammonium sulphide, yields a non-reducing substaiice with acid 
properties, not the original dextrin from which we started. We have 
found, however, that the dextrinic acid which is obtained by the 
oxidation of the dextrin with mercuric oxide and baryta is itself 
capable of nitration, and that the dextrinic acid can be readily 
regenerated from the nitrate in the usual manner, and with all its 
properties unchanged. 

By this process, we are able to obtain the dextrinic acid quite free 
from ashy and consequently in a state suitable for the examination of 
its salts. 

* The standard time of hydrolysis of 14 hours was not found sufficioDt to com- 
pletely hydrolyse the dextrinic acid, which required 24 hours treatment under the 
conditions described in the Appendix to our paper on maltodextrin (this vol., p. 805). 

Digitized by VjOOQIC 


A specimen of dextrinic acid with an optical activity of [a]D 193^, 
and yielding 105*9 per cent, of glucose on complete hydrolysis with 
oxalic acid, was obtained as a thick syrup on concentrating its solu- 
tion to the utmost on the water-bath. It was then mixed in the cold 
with ten times its weight of strong nitric acid, sp. gr. 1*4, and the 
nitrate thus formed was precipitated by strong sulphuric acid and sub- 
sequent dilution with water in the usual manner, the temperature 
being maintained as low as possible by immersing the vessel in ice 
cold water. The nitrate, after being well washed with water, was 
dissolved in glacial acetic acid and filtered ; water added to this 
solution again threw out the nitrate, which was once more dissolved 
in acetic acid and reprecipitated. This operation was twice repeated, 
and the nitrate was then treated with ammonium sulphide; the 
recovered dextrinic acid was completely freed from ammonium salts 
by repeated precipitations with alcohol of 95 per cent., seven or eight 
such treatments generally being necessary before the Nessler test ceased 
to give a reaction. It was finally obtained in a state of purity, and 
quite free from ash, as a white, amorphous substance having feeble 
but distinct acid properties, and with a specific rotatory power cor- 
responding to [a]D 193-7° and R against [a]© 193*0° of the original 
acid before nitration. After completely drying in a vacuum at 100° 
over phosphoric anhydride, its divisor for a solution density of 
1040*36 was found to be 4005. 

On complete hydrolysis with oxalic acid, it yielded 106*2 per cent, of 
glucose, against 105*9 given by the original dextrinic acid before 

This experiment is an important one, not only in showing that the 
product recovered from the nitrated dextrinic acid is identical with 
the dextrinic acid before nitration, but also in showing that the acid 
properties of the oxidised dextrin cannot be due to the accidental ad- 
mixture of other acids produced by the oxidation of a reducing carbo- 
hydrate present in the original dextrin. Were this the case, we 
should have to assume that the acids proceeding from this reducing 
carbohydrate not only persistently adhere to the dextrin during the 
repeated fractionations of dextrinic acid itself, but that even the 
process of nitration, fi*actionation of the nitrate, and the subsequent 
processes of recovery have also been incapable of effecting differentia- 

In the face of all this evidence, we think it can no longer be 
doubted that the carboxylic constituent of the oxidised dextrin really 
forms part of the molecular complex constituting dextrinic acid. 
To admU this fact m, qf cotiTBe, to admit that the aidehydic or ketonic 
portion qf the original dextrin, to which it owes its reducing power oj 
B 6'5,is also part of the moleculctr complex of the stable dextrin, t 

_. , ^oogle 


VI. Analysis of the Calcium Salt of Dextrinie Acid, 

We have already stated that dextrinie acid possesses only very 
feeble, although well-marked, acid properties. It readily forms 
barium and calcium salts on boiling the acid with the respective 
carbonates, and these salts can be precipitated from the solution by 
dilute alcohol. They are readily soluble in water, but like the cor- 
responding derivatives of the maltodextrinic acids, they cannot be 
obtained in the crystallised form. 

Owing to its complexity and feeble basicity, the dextrinie acid must 
be abaoltUely free from ash before it can be used for the preparation 
of its salts, otherwise their composition could not be even approxi- 
mately determined. 

The process of nitration of the acid and its subsequent regeneration 
from the nitrate fortunately enables us to prepare the dextrinie acid 
in this state and in a high degree of purity. 

An analysis of the calcium salt after most careful preparation in 
this manner shows that it contains 0*29 per cent, of calcium. 

YII. Hyd/rolyeia of Deoctrinic Add unih OaxUie Acid and Diastase, 

We have already seen that when free dextrinie acid is completely 
hydrolysed with oxalic acid, the main product is glucose, which 
amounted in one experiment to 105*9 per cent, and in another to 
106*2 per cent. The hydrolysis of the calcium salt in a similar 
manner was found to produce 107*2 per cent, of glucose on the dry 

It was always found in such experiments that, besides glucose, the 
complete hydrolysis gave rise to a small quantity of an acid of low 
molecular weight, the production of this fuaal acid being strictly 
analogous to that of the C^-acid when the maltodextrinic acids are 
hydrolysed, or to that of gluconic acid when maltobionic acid or its 
salts are treated with dilute acids. 

The large yield of glucose in the case of the dextrinie acid 
hydrolysis clearly shows that the carboxyl portion of the original 
complex must bear a very small relation to the total molecule, a 
result which is also quite in accord with the small reducing power 
of the dextrin from which the dextrinie acid has been derived. 

Since we could not expect more than a 2 '5 per cent, yield of the 
final acid, it was necessary to hydrolyse a considerable quantity of the 
dextrinie acid in order to obtain a suficient quantity of this sub- 
stance to admit of a study of its properties. 

In order to see if it were gluconic acid, the mixed products from 
the hydrolysis were fermented with yeast, and the residue, after 

Digitized by VjOOQIC 


evaporation and treatment with calcium carbonate, was evaporated 
with alcohol and fractionated in such a manner as to separate any 
trace of calcium gluconate. No evidence of its existence could be 
obtained, although it had been shown in preliminary experiments 
that even an extremely small quantity of previously added calcium 
gluconate could be recovered in a crystalline form by the method 

A sufficient quantity of the final acid for complete examination was 
prepared by completely hydrolysing 29 grams of dextrinic acid by 
heating it with a 2*5 per cent, solution of oxalic acid for 23 hours 
at 100^. The solution of the hydrolysed products, after treatment 
with calcium carbonate, was filtered and evaporated to the thickest 
possible syrup, which was then repeatedly extracted with hot 95 per 
cent, alcohol to remove the glucose.* 

The residual calcium salt was redissol ved in a small quantity of water, 
filtered, and again thrown down with alcohol ; at this stage, the salt 
which was evidently still impure, contained 9*0 per cent, of calcium. 
After several further treatments of a similar kind, and a partial frac- 
tionation with alcohol, the salt was obtained with constant properties. 
The following results were obtained on analysis. 

Calculated for {C^UgO^\G& .. 32-4 
Found 32-1 

The salt corresponds in composition and general properties with the 
final Cg-acid which we obtained by complete acid hydrolysis of the 
two maltodextrinic acids, and also with the acid obtained by the 
oxidation of maltose with mercuric oxide and subsequent hydrolysis 
of the product (this vol., p. 296). 

Dextrinic acid and its salts, unlike maltodextrinic acid A, are but 
very slowly influenced by diastase. A marked action cau, however, 
be detected if the digestion with malt extract in the cold is allowed 
to go on for several days, or if the amount of diastase used is 
relatively large, and the action is carried on for 48 hours or so at 

The product of hydrolysis in this case is a mixtwe of mcUtoae cmd 
glucose, a fact which will be further considered in the next section in 
connection with the diastase-hydrolysis of the parent dextrin itself. 

* The glucose waa cryatalliaed out from these alcoholic extracts. It proved to be 










Digitized by VjOOQIC 


YIII. HydrolysU of the Stable Deoafyrin vnth OxeUie Acid and 
fvUh Diastase, 

When the stable dextrin is hydrolysed with a 2 per cent, solution 
of oxalic acid under the standard conditions, it is completely converted 
into d-glucoae. We here give the result of such an experiment, the 
glucose being determined by two independent methods. 

Glacose from 100 parts 
of dextrin. 

1 . — Estimated by increase of sp. gr , . . 1 1 0*8 

2. — „ from cupric reduction 108*9 

The well-marked ' resting stage ' of the products of starch trans- 
formations brought about by diastase, when these products have 
reached an [aJD of 150° and B 80, has already prepared us for the 
fact that the dextrin, when isolated, has a considerable resistance to 
the hydrolytic influence of diastase. The observed difference in the 
rapidity of this action on the dextrin, on the one hand, and on malto- 
dextrin or any of the intermediate transformation products, on the 
other, is very striking. Whereas solutions of maltodextrin of a con- 
centration of from 5 to 10 per cent, are completely resolvable into 
maltose by about half an hour's digestion at 50° with an amount of 
malt extract corresponding to but 2 or 3 c.c. per 100 c.c. of solution, 
a dextrin solution requires several hours of similar treatment with 
ten times the amount of hydrolysing agent before even an appreciable 
amount of change is produced. 

We have made this experiment several times, with every necessary 
correction, and have found that a solution of the stable dextrin of 
7 per cent, concentration, with the addition of malt extract at the 
rate of 25 c.c. per 100 c.c. of solution, is only hydrolysed to the extent 
of about 30 per cent, after 48 hours at 55°. 

The products of hydrolysis in this case consisted of almost exacUy 
equal qtuxntities qf maltose and deostrose, a fact which was shown, not 
only by the relation of the optical and the reducing properties of the 
products, but also by the preparation, separation, and estimation of 
the osazones. 

At first sight, it might be considered that the glucose found in this 
reaction was a secondary product, the dextrin yielding in the first 
place maltose, which was subsequently hydrolysed to glucose by the 
enzymes of the malt-extract. This explanation might certainly 
apply in those cases where an air-dried malt is employed, for an 
extract of such a malt often contains a transforming agent which, by 
long continued action, has an appreciable effect in hydrolysing maltose. 
In this case, however, a kiln-dried malt was employed whose [aqueous 

„ .gitized by VjOOQ _ _ 


extract was shown to be incapable of acting on maltose, even after 
60 hours digestion at 55°. We must, therefore, regard the formation 
of maltose and glucose during the hydrolysis of the dextrin as being 
sifnultaneatu and not stieoesnve phenomena, a fact which, in con- 
junction with their approximately equal production, is of some 
importance in its bearing on the constitution of the dextrin. 

IX. GeTieralisatiana and Conduaions, 

When starch is transformed by an active diastase, such as that 
derived from an air-dried malt, at temperatures below 60°, the reaction 
proceeds very rapidly until a resting stage is reached corresponding 
to [a]], 150° and R 80 for the mixed products of change. At this 
point, these products consist of maltose, and a dextrin which, com- 
paratively speaking, is very resistant to the further action of the 
hydrolytic agent. This dextrin has a specific rotatory power of 
[a]]) 197 — 198° and, even when obtained in the purest possible state, 
has a feeble reducing power corresponding to B 5'5. 

Under the mild oxidising influence of mercuric oxide and baryta, 
the dextrin is converted into a complex carboxylic acid — dextrinic acid. 

Under the hydrolytic action of dilute acids; this carboxylic 
polysaccharide is resolved into df-glucose and a residual Og-acid which 
is identical with the CgH^QO^-acid obtained either from the acid 
hydrolysis of the maltodextrinic acids, or from the hydrolysis of the 
biose acid proceeding from the oxidation of maltose with mercuric 

In the case of maltodextrin and the maltodextrinic acids, on the 
other hand, the complex is certainly resolvable by diastase into 
Cj2-groups only, that is, into maltose. 

At first sight, this points to there being a difference in the linkings 
or spatial relations between the O^^'S^^^P^ ^^^ their constituent 
C^-groups, and since in the case of dextrin and its derivative dextrinic 
acid there is no selective disposition on the part of diastase to split 
off maltose rather than glucose, we are justified in regarding the 
stable dextrin as made up of Cg-complexes.* 

A consideration of all the facts we have brought forward justifies 
us in regarding the stable dextrin, from an empirical point of view, 
as a combination of 39 groups of CgHi^Og- and one C^HuO^-group or 
as40(CeHioO,) + H20. 

* In pushing this explanation to its ultimate conclusion, we are, however, met 
with the diflaculty that the diastase of the kiln-dried malt used in these experi- 
ments is incapable of hydrolysing maltose. One would expect that a hydrolysing 
agent which could determine the resolution of the dextrin complex either into C^- 
or Ci2-gro^P> would also be able to resolve any of the fission products into glucose, 
but Uiis apparently is not the case. 

Digitized by VjOOQ IC 


Its constitution may, however, be more rationally regarded as a 
condensation of 40 glucose molecules with the elimination of the 
elements of 39 molecules of water. 

This constitution may be expressed in the following manner, only 
one of the 38 similar intermediate groups being represented for the 
sake of brevity. 



0<;(CH0H)2 I 0<^(CH0H)2 I 


38 times repeated. 

On this view, the constitution of dextrinic acid will be given by 
converting the terminal right hand group to -CH2'[CH'OH]3«COOH, 
which is the residue of the O^H^QOg-acid given by the complete 
hydrolysis of dextrinic acid with acids. 

These complicated constitutional formulse can be written in an 
abbreviated form somewhat similar to that employed for the con- 
densed formula of maltodextrin. 

0-C,HioO, 0-CeHioO,. 

stable dextrin. Dextrinic acid. 

The oxygen atoms on the left in each case represent the 39 oxygen 
Unkings between the 40 groups. The sign <[ is again used to 
indicate the open carbonyl of the terminal group. 

We have now to see how far the above views are in consonance 
with the observed facts. 

It is manifest that an ultimate analysis of such substances cannot 
determine their exact composition. It would, in fact, be quite 
impossible by a mere combustion to differentiate a dextrin made up 
of 39 groups of OgH^QOg- and one C^HjgOg-group, from one wholly 
made up of C^H^QO^-groups, since the carbon percentages are within 
0*12 per cent. It is only by such methods as we have described, that 
is, by examination of the salts of the derivative acids, and a quan- 
titative study of the products of hydrolysis, that we can expect to 
have any light thrown on their approximate constitution. 

Commencing with the dextrinic acid, the calcium salt of an acid of 
the above constitution should contain 0*30 per cent, of calcium ; the 
actual amount of calcium found in the pure salt was 0*29 per cent. 
The amount of glucose which, on the above view, ought to be obtain- 

Digitized by VjOOQIC 


able from free dextrinic acid by complete hydrolysis is 108*2 per cent., 
and 107 '9 per cent, from its calcium salt. The glucose obtained in 
our experiments from the free acid was in one case 105 '9 per cent., and 
in another 106*2 per cent., whilst the dry calcium salt of the 
dextrinic acid gave 107*2 per cent. 

Turning now to the acid hydrolysis of the dextrin itself, we note 
that, with the constitution which we have attributed to it, it ought to 
yield on complete acid hydrolysis 110*8 per cent, of glucose. The 
actual amount found by one process was 110*8, and by another 108*9, 
the error in the second determination being certainly a mtTwa one. 

We have seen in a previous paper how the cupric-reducing power of 
the maltose residue of maltodextrin is approximately that of an 
equivalent quantity of maltose. In the same way, it appears that the 
glucose residue, constituting the reducing portion of stable dextrin, 
also retains the reducing power proper to dextrose, for if we regard, 
as we are entitled to do from the study of its oxidation products, the 
stable dextrin as constituted of 39 CgH^^Og groups .and one glucose 
group, the ' apparent glucose ' in this complex, as determined from the 
actual amount of cupric oxide it reduces, is about 2*7 per cent. This 
corresponds to an apparent maltose percentage of 5*5, that is, to an 
B of 5*5, which is exactly the value expressing the reducing power of 
the dextrin in a pure state. 

It now remains to discuss the bearing of our recent work on the 
size -'and constitution of the molecule of starch, or rather that of 
soluble starch, which differs materially in physical properties from the 
substance forming the greater part of the starch granule. 

The commonly accepted empirical formula for soluble starch, 
nC^H^QOg, or nC^^^QO^Q, is doubtless correct, since it is not only in 
accord with numerous combustions made from time to time by different 
observers, but is also fully confirmed by our own work, described in a 
previous paper (this vol., p. 307), in which we have accurately deter- 
mined the amount of glucose which it yields under acid hydrolysis, 
and also the actual amount of products obtained by diastase- 

In our previous writings (Brown and Heron, Brown and Morris, &c.), 
we have frequently drawn attention to the fact that the quantitative 
relations of the maltose and dextrin, occurring in starch transforma- 
tions which have been brought down to the 'resting stage' by 
diastase, are in close agreement with the view that the reaction 
takes place according to the empirical equation represented by 

lOOi^oOio + 8H,0 = SCijHjjOii + 2G^^JI^0,^. 
Starch. Maltose. Stable dextrin. 

This is the ' No. 8 equation ' of our earlier writings. It assumes 

VOL. LXXV. „gitizedbyG£)Ogle 


that the dextrin is a substance without cupric-reducing power, and 
with an empirical formula of nC-^JB.2Q^i(fi ^^ ^^e^io^s* -^^ ^® ^^^ 
know that the stable dextrin is not altogether free from reducing 
power, and that its formula cannot be expressed in a simpler form 
than by ^OCqB.^qO^ + 'EL^O, the above equation requires modifying in 
such a manner as to show the interaction of ^ mol. H^O more than the 
eight molecules indicated. The multiplication of both sides of the 
equation by 10, and the introduction of 1 mol. H2O more, brings the 
theoretical expression in close accord with the facts : thus 

Starch. Maltose. Stable dextrin. 

This requires the mixed products of transformation at the close of 
the reaction to have optical and reducing properties corresponding 
to [a]i> 149*2% B 81*8, whilst the old uncorrected expression requires 
[ajo 150'2% B 80*8. The experimental values are always very dose 
to[a]Dl50%B80— 81. 

It must be borne in mind that the value of .B 81*8 cannot now be 
regarded as being quite a measure of the actual amount of maltose 
present in the products of one of these low transformations, owing to 
the fact of the dextrin being slightly reducing. The actual amount 
of free fermentable maltose corresponds to 80*8 per cent., the 
difEerence of 1 per cent, being accounted for by the reducing power of 
the dextrin. 

From the quantitative relations of the maltose and dextrin found 
in the mixed products of starch hydrolysis which have been brought 
down to the resting stage, it is clear that the size of the starch 
molecule cannot be less thsknfive iimeg that of the stable dextrin. 

The study of the dextrinic acid derived from the dextrin has given 
us a minimum expression of the size of the dextrin molecule, ap- 
parently represented by ^9{C^'Bi^QO^),Of^K^fi^9 which indicates a 
molecular weight of 6498. 

In the year 1889 (Trans., 55, 462), one of us, in conjunction with 
G. EL Morris, described certain attempts which we had made to 
determine the molecular weight of some of the starch derivatives by 
means of the freezing-point method, and amongst others we examined 
the stable dextrin prepared from a starch transformation with 
diastase run down to the ' resting stage.' This dextrin, prepared by 
alcoholic precipitation, had optical and reducing properties correspond- 
ing to [a]D 196*2% B 7*3. Its molecular weight, deduced from the 
mean of a large number of concordant experiments made on the 
depression of the freezing point of a strong aqueous solution, was 6221. 
Although we are less disposed than formerly to base any arguments 

Digitized by VjOOQ _ _ 

thansfobmations, and relation to maltodextrins, etc. 335 

on the molecular weight determinations of these colloids by the 
freezing-point method, it is certainly remarkable that we should have 
been led to practically the same conclusion as to the molecular weight 
of the dextrin by considerations entirely independent of each other. 

Granted that the molecular weight of the stable dextrin is 6498, 
the complex representing the molecule of soluble starch cannot have a 
less molecular weight than 32,400, its empirical formula being repre- 
sented by lOOCigHgoOjo or^by (80Ci2H2oOio,40CeHio05). The latter 
expression connotes the fact that the one-fifth of the molecule which 
gives rise to the stable dextrin on hydrolysis, differs from the 
remainder which is directly hydrolysable to maltose. The known 
properties of soluble starch would in themselves suggest a very high 
molecular weight ; its highly colloidal nature, the ease with which it 
can be removed from its solutions by merely forcing them through 
porous earthenware, and the extremely small influence it exerts on the 
freezing point* of its solvent all pointing to a very high molecular 
complexity, a complexity which probably approaches that of some of 
the proteids. 

Since starch is a non-redacing polysaccharide it does not contain 
free carbonyl groups. 

The simplest manner in which it can be expressed constitutionally, 
with due regard to all the facts, is to consider it as made up of the 
residues of 80 mcUtan groups and 40 dextran groups, linked in ring 
form through oxygen atoms. 

On hydrolysis, the dextrcm complex, constituting one-fifth of the 
whole molecule, is split off with the formation of the stable dextrin, 
SOCgHi^OgjCgHi^Og, whilst the maltan portion of the ring is attacked 
at the oxygen linkings of the Cj2'^^^P^> ^^^ hydrogen ions of the 
reacting water molecules moving in one direction, and the hydr- 
oxyl ions in the other, thus forming, by successive stages of hydrolysis, 
maltodextrins or ** reducing dextrins,"t and finally maltose4 

We here give a condensed representation of our views on the 
constitution of the molecule of soluble starch, A representing the 
dextran residue, and B the four equal maltan residues. Although in 

* We found in 1889 (see Brown and Morris, Trans., 55» 465) that the depression 
of the freezing point of an aqueous solution of soluble starch was extremely small, 
indicating a molecular weight between 20,000 and 30,000. 

t The term mdUodextrin is preferable to that of ' redvd'ng dextrin,* since the 
latter does not connote the all-important fact that such substances do not contain 
any stable dextrin residue, but are wholly convertible into maltose by diastase 

t The manner in which the activity of the diastase has been previously modified 
by. heat, the time of the reaction, and the temperature at which it takes place, are the 
principal factors in determining the amount of degradation of the maltan portion of 
the starch. ^ j 

Digitized byVjOOQlC 


this condensed constitutional formula we indicate the maltan con- 
stituents of the molecule as being airranged in /our groups, this is not 
essential to the hypothesis here put forward, for it is possible that 
the oxygen linkings between the individual C^j-groups may all be of 

o o o 

V°v / 


o 'o 

* It will be observed that there are some points of similarity between this 
constitutional formula for starch and that suggested by Dr. Armstrong in 1880 
(see Armstrong's Miller's EleinerUs, p. 658). 

the same order, and that the fission due to hydrolysis may occur 
indifferently at any point of the maltan portion of the chain. The 
experiments, however, which we described in 1889 (Trans., 56, 469) 
on the freezing of solutions of high starch products and of trans- 
formations arrested at an early stage of hydrolysis, certainly suggest 
that the early fission products have a molecular complexity about 
equal to that of the stable dextrin itself ; in other words, that the first 
act of hydrolysis is to divide the starch molecule into five approxi- 
mately equal parts. 

Whilst a consideration of the new facts we have brought forward 
in these recent papers has led us to very much the same view as that 

„ .gitized by VjOOQ _ _ 


held by one of us and Morris ten years ago as regards the size and 
complexity of the starch molecule, our study of the whole question 
during the last four years has enabled us to substitute for the tenta- 
tive hypothesis then put forward one which is perhaps a step nearer 
the truth, and at any rate has the advantage of being more in conson- 
ance with the ordinarily accepted views of the constitution of the 

According to the old view, we pictured the starch molecule as 
consisting of four complex amylin groups arranged round a fifth 
similar group constituting the portion of the complex which ulti- 
mately became the stable dextrin. It was considered that the first 
action of the diastase was to liberate these five complexes in the form 
of non-reducing dextrins, and that four of these differed from the 
fifth in being hydrolysable by successive gradations down to maltose. 
At that time, the existence of high, degradable, non-reducing dextrins 
was considered probable, but further investigation has convinced us 
that such substances do not exist as separate entities, and that all 
the fission products of starch hydrolysis are, to some extent, reducing, 
the extent of this reducing power being dependent on the relation of 
the end carbonyl group to the size of the complex to which it is 

Davy-Faraday Research Laboratory, 
Royal Institution. 

XXXIIL — Position Isomerism and Optical Activity ; the 
Methylic and Ethylic Salts of Benzoyl and of Oriho-, 
Meta-, and Para-malic Acid. 

By Pebot Fbanklakd, Ph.D., F.It.S., and Fbedebick Malcolm 
Wharton, AI.C., late Priestley Scholar in Mason University 
College, Birmingham. 

The present investigation forms a continuation of the work on the 
same subject which we have already published in the pages of the 
Journal of the Chemical Society (Trans., 1896, 69, 1309—1321; 
1583—1592; also P. Frankland and McCrae, Trans., 1898, 73, 
307 — 329) and which has for its object the determination of the 
influence respectively exerted by position-isomeric groups attached 
to the asymmetric carbon atom. 

Beasons have been adduced by us for anticipating that, in the case 
of isomeric di-substituted benzene rings, the para-isomeride should 
exert the greatest, the ortho- the least, and the meta- an intermediate 

„.gitized by Google 


influence on the rotation, and experiment shows that this sequence is 
80 far, with two doubtful exceptions (Trans., 1898, 73, 308 and 309), 
uniformly maintained. We have also given reasons for anticipating 
that the unsubstituted phenyl group should exert a rotatory influence 
inferior to that of the para- or meta*groups, but according to circum- 
stances either greater or less than that of the ortho-groups. Experience 
shows, however, that the rotatory influence of the unsubstituted 
phenyl group is apparently much more erratic, occupying every 
possible position as regards the magnitude of the rotatory influences 
of the three isomeric di-substituted benzene rings. 

Preparation of Ethylic and Methylic Malates. 

The difficulties attending the preparation of the ethereal salts of 
malic acid have been pointed out by Anschatz and Beitter (Zeii, 
physikaL Chem,, 1895, 16, 493), who have shown that if the ordinary 
hydrochloric acid method is employed, a considerable proportion of the 
malic acid is converted into f nmaric acid unless a very low tempera- 
ture is preserved during the saturation of the alcoholic solution of 
malic acid with hydrogen chloride. It has also been shown by 
Purdie, Williamson and Lander (Trans., 09, 829; Purdie and 
Lander, Trans., 73, 287) that if the method of acting on silver malate 
with an alkyl iodide is employed, the alkyl malate is mixed with the 
ethereal salt of the corresponding alkoxysuccinic acid which entirely 
vitiates the optical activity of the product. 

In choosing between these two evils, we have resorted to the hydro- 
chloric acid method as the less objectionable of the two, and we 
have carefully observed the precautions indicated by AnschUtz so as 
to secure the maximum purity of the product. 

Ethylic Malate. — Fifty grams of finely powdered malic acid(Elahlbaum), 
dried in a vacuum desiccator for 3 days, were mixed with 100 grams 
of absolute ethylic alcohol in which the acid nearly all dissolved ; a 
slow stream of carefully dried hydrogen chloride was then passed through 
the liquid, which was kept at - 18° by means of a freezing mixture. 
When completely saturated, the .liquid was left for 24 hours, a 
current of dry air was then drawn through for 48 hours, after which it 
was placed in a vacuum desiccator containing slaked lime, until the 
smell of hydrochloric acid had nearly disappeared. The excess of 
alcohol was then distilled o£E under reduced pressure and at as low a 
temperature as possible, the ethereal salt being subsequently distilled 
on an oil-bath. Fifty-one grams of the crude ethereal salt were ob- 
tained between 128° and 140° (11 mm. pressure). This exhibited 
the rotation 

ap=- 10-30°; /=!, < = 21° 

Digitized by VjOOQIC 


This was redistilled^ the portion passing over at 129 — 134^ 
(12 mm. press.) having a rotation 

od=-11-56°; ^ = 1, < = 17-6^. 

After treatment in ethereal solution with fused potassium car- 
bonate and redistillation until the rotation was constant, it finally 
distilled at 129 — 132° (11 mm. pressure), and exhibited the specific 

t°]y- lo"x 11340 ° -^'^•'**° <** 2074° = 11340), 

a value agreeing closely with that obtained by Purdie and William- 
son, namely, -10*30°, from Elahlbaum's malic acid by use of the 
hydrochloric acid method. 

0-2302 gave 0-4229 OOj, 0-1510 HjO, C « 5010 ; H = 7-29. 
0-2083 „ 0-3834 COj, 01368 HgO. = 5020 ; H -r 7-30. 
CgHi^Og requires = 5053 ; H = 737 per cent. 

Methylie Maiate, — ^This was prepared in exactly the same way as the 
ethylic compound. Forty-eight grams of crude ethereal salt, boiling 
at 128 — 142° under 12 mm. pressure, were obtained from 50 grams of 
dry malic acid and 100 grams of methylic alcohol. This product had 
a rotation 

a,> -.-6-70°; Z=l,<-18-6°. 

The product was fractionated under diminished pressure until of 
practically constant rotation; the final product boiled at 129° 
(11 mm.), and exhibited the following rotations. 

ai>= -8-36°; ^=1, < = 16-6° 
a„= -8-41 ; Z=l, < = 20 
aD=-8-46 ; ^=1,« = 40 

WJT- ix 1-2301 =-*'®*°- 

Purdie and Williamson give as the observed rotation ax>» - 8*17°. 
Anschiitz and Reitter [ a J, = - 6-88° Walden [a\, = - 6-85°. 

0-2300 gave 0-3735 COj and 01 285 Hfi, C = 44-29 ; H«6-21. 
0-2191 „ 0-355100, „ 01212 HgO. C-44-20;H = 615. 
CgHj^Oj requires C = 4444 ; H = 617 per cent. 

Ethylic BenzoylmalaU, 

Ethylic malate was slowly added to an excess of benzoic chloride 
heated in an oil-bath at 140°, the temperature being subsequently 
raised to 170°, at which it was maintained for 40 minutes. The excess 
of benzoic chloride was distilled off first i;nder reduced pressure and 

_. , ^oogle 


then, on raising the temperature, the crude ethyiic benzoylmalate 
passed over at 210 — 220° (12 mm.) ; the latter was refractionated 
under diminished pressure until the rotation was constant. It was a 
thick, colourless liquid which we were unable to obtain in the solid 


0-2210 gave 0-4941 CO3 and 01 222 H^O. = 60-97 ; H = 6-14. 
0-2153 „ 0-4818 COg „ 01203 H^O. = 61-03; H = 6-21. 
OigHigOe requires = 61-22 ; H* 6-12 per cent. 

<i 4074°= 11361; c^60°/4°= 1-1253; £^60^/4*^= 1-1158. 

With this product, the following polarimetric determinations were 

Rotation of ethyUc benzoylmalate. 

Obaerved rotation an 

Density compared 


in 99-9 mm. tube. 

with water at 4°. 



- 4-47° 


- 3-87° 


- 6-02 ■ 




















- 10-57 




- 12-48 


- 12-08 

Methylic Benzoylmalate* 

The mode of preparation was similar to that described above for 
the ethyiic compound. The crude product was dissolved in chloro- 
form, well shaken with a solution of sodium carbonate, and, after 
separation, the chloroform solution was dried with fused potassium 
carbonate. After removing the chloroform, the liquid was fraction- 
ated under reduced pressure until the rotation was constant, it passed 
over at 210 — 223 (12 mm.) ; it could not be induced to solidify. 

0-2346 gave 0-4991 OOg and 01147 HjO. = 58-02 ; H = 5-43. 

0-2312 „ 0-4951 00, „ 0-1105 HgO. = 58-40 ; H = 5-31. 

0-2242 „ 0-4798 OOg „ 0-1082 HgO. = 58-37 ; H = 5-36. 

OigHj^Og requires = 58*65 ; H = 5-26 per cent. 

The following density determinations were made. 

d 4074°= 1-1944; d 6074°=M759 ; d 7074°= M665, 

and the following polarimetric observations. 

Digitized by VjOOQIC 


notation of methylie benzof/lmalate. 


Obearred rotation ad 
in. 99 '9 mm. tabe. 

Density compared 
with water at 4°. 















- 10-93 




- 13-16 






- 13-64 

Ethylic Orthotoluylmaiate, 

This was prepared in the usual manner by adding ethylic malate 
to twice the theoretical quantity of orthotoluic chloride heated at 
140^ and ultimately to 170°. The excess of acid chloride was distilled 
off under reduced pressure, the residue being then dissolved in 
chloroform and shaken for 24 hours with a solution of sodium 
carbonate to remove any remaining acid chloride. After washing 
repeatedly with water, the chloroform solution was dried with 
potassium carbonate, and the chloroform distilled off. The rotation 
of the crude product thus obtained was 

It was purified by repeated fractionation under reduced pressure 
until the rotation was constant ; it distils at 215 — 225° (12 mm.). 

0-2170 gave 0-4895 QO^ and 01277 HjO. - 61-52 ; H = 6-54. 

0-2072 „ 0-4719 COj „ 0-1227 HgO. = 6211 ; H- 6-58. 

0-2088 „ 0-4759 COj „ 0-1231 HgO. = 62-16 ; H = 6-55. 

CjgHjoOe requires = 62-34 ; H = 6-49 per cent. 

The following density determinations were made, 

cl40°/4° = 1-1228; d 6074°= 1 -1057; d 7074°= 1-0968, 

and the following polarimetric observations. 

Rotation of ethylic orthotoluylmcdate. 


Observed rotation aj, 
in 92*36 mm. tube. 

Density comi>ared 
with water at 4*. 




- 6-58° 














- 10-38 





Digitized by Google 


Metkylio Orthotoluf/lmalaie, 

This was prepared in a similar manner to the ethylio salt. The 
crude compound, after washing with sodium carbonate solution, 
exhibited the following rotations. 

aD= - 8-56° ;(^=1,« = 9-5°) 
aD= -.1308^; (^=1,^ = 100°) 

As solidification could not be induced, the purification was effected 
by fractionation under reduced pressure until the rotation was 
constant ; the liquid passed over at 214 — 225° (12 mm.). 

0-2174 gave 0-4759 00^ and 0-1123 HjO. = 59-70 ; H = 5-74. 
0-2216,, 0-4841 COj „ 0-1131 HjO. = 59-58 ; H-5-67. 
0-2307,, 0-5054 OO2 „ 0-1196 H2O. = 59-75; H = 5-76. 
Oj^HijOe requires = 60-00; H = 5-71 per cent. 

The following density determinations were made, 

<i 4074°= 1-1744; c? 5074° =1-1647; d 6074°= 1-1550, 

and the following polarimetric observations. 

Eolation of methylic ortkotoluylmaiaie. 

Observed rotation ao 

Density compared 


in 99*9 mm. tube. 

with water at 4°. 



- 10-61° 


- 8-94° 








- 10-06 


- 12-45 


- 10-83 




- 12-40 


- 16-23 



Ethylio Metaioluylmalate, 

The preparation was effected in exactly the same manner as that 
already described for the orthotoluyl compound. The product was 
a colourless liquid distilling at 212—220° (13 mm.). 

0-2389 gave 0-5435 OO2 and 01427 HgO. = 6205; H = 6-64. 

0-2328 „ 0-5290 002 „ 0-1383 HgO. = 61-97; H = 6-60. 

OieHjoOg requires = 62-34 ; H = 6*49 per cent. 

The following density determinations were made, 

(i40°/4° = 1-1185; (i50°/4°= 1-1086 ; rf 60°/4° = 1 -0989, 

and the following polarimetric observations. 

Digitized by VjOOQIC 


RoUUion of ethylic metcUoluylmalata, 
Observed rotation an Density compared 


in 99-9 mm. tube. 

with water at 4'. 




1-1371 ' 






































Methylic MetatoluylmcUate, 

This was obtained in like manner, as a thick, almost colourless, liquid 
distilling at 215—225° (12 mm.). 

0-2094 gave 0*4599 COg and 01098 Kfi. C = 59-90; H = 6-83. 

0-2126 „ 0-4665 COa „ 0-1108 H2O. = 59-84; H = 5-79. 

Ci^HigOg requires = 6000 ; H = 5-71 per cent. 

The following density determinations were made, 

d 4074° =1-1723; d 50714''« 1-1622; d 60°/4°= 11521, 

and the following polarimetric observations. 

notation qf methylic metatoluylmcdate. 
Obser?ed rotation ap Density compared 


in 99-9 mm. tube. 

with water at 4°. 





- 6-34° 














- 13-04 




- 14-49 


- 13-49 

Ethylic ParcUoluylmalate. 

The preparation was carried out in exactly the same manner as 
already described for the orthotoluyl compound. The ethylic salt 
was obtained as a thick, colourless liquid which was purified by 
fractionation under reduced pressure until the rotation was 

Digitized by VjOOQIC 


0-2253 gave 0-6111 OO2 and 0-1320 HgO. = 61-87 ; H = 6-51. 

0-2215 „ 0-5021 CO2 „ 01304 H2O. = 61-82; H = 6-54. 

OigHjjjOg requires 0« 62*34 ; H = 6-49 per cent. 

The following density determinatations were made, 

d 4074°= 1-1162; d 6074°= 1-1052; d 6074^=1-0943, 

and the following polarimetric observations. 

Rotation qf ethylic paratoluylmalate. 

Observed rotation a D Density compared 

Temperature, in 99*9 mm. tube. with water at 4°. [a\o. 

20^ -0-25° M382 -0*22° 

29 -1-06 M283 -0*94 

35 -1-71 1-1217 -1-53 

51 -2-99 1-1041 -2-71 

62 -3-71 1-0920 -3-40 

99 -5-83 1-0513 -555 

136 -7-38 1-0106 -7-31 

Note, — On the first occasion when we prepared ethylic paratoluyl- 
malate, this substance, before distillation, but after it had been washed 
in chloroform solution with sodium carbonate, became quite thick 
through separation of crystals ; the latter were recrystallised several 
times from methylated spirit until of constant melting point (95°), 
when they presented the appearance of long, shining blades. These 
crystals' dissolved slowly in caustic potash on heating, and on adding 
hydrochloric acid, paratoluic acid melting at 176° was deposited. 
On analysis, 

0-1549 gave 0-4257 OOg and 00787 H^O. = 74-95 ; H = 5-65. 
0-1812 „ 0-4990 OO2 „ 00919 HjjO. = 7511 ; H = 5-64. 

These figures negatived the possibility of this substance being 
ethylic paratoluylmalr. 'c, but, on the other hand, they point unmis- 
takably to its being ; cratohiio anhydride, Oi^Hj^Og, which requires 
= 75-60, H = 5-51 p.r cent. As this compound does not appear to 
have been prepared before, we had no further means of identifying it 
with the quantity at our disposal. Benzoic anhydride melts at 42°. 

Meikylic ParatoluylTtutUUe, 

Even this ethereal salt was only obtained as a liquid distilling 
between 200° and 225° under 13 mm. pressure. 

0-2317 gave 0-5074 OOg and 0-1213 HgO. =59-72 ; H = 5-82. 
0-2264 „ 0-4955 OO2 „ 01192 HgO. = 69-69 ; H« 5-86. * 
^14^16^6 requires O--6000 ; H = 5-71 per cent. 

„.gitized by Google 


The following density determinations were made, 

d 40°/4°=M725 ; d 5074°=M617 ; d 60°/4°=11509, 
and the following polarimetric observations. 

Rotation of methylic poflratoluylnudate, 
Obseired rotation an Density compared 


in 99 '9 mm. tobe. 

with water at 4°. 



- 3-75° 























136 -10-83 1-0688 -10-14 

The results recorded above may be summarised as follows. 

1. The IsBVorotation of ethylic is greater than that of methylic 

2. The Isevorotation of ethylic malate is diminished by the intro- 
duction of the benzoyl and three toluyl groups. The laevorotation 
of methylic malate is also diminished by the introduction of the 
benzoyl, paratoluyl, and metatoluyl groups, but is increased by the 
introduction of the orthotoluyl group. The exceptional influence of 
the orthotoluyl group on methylic malate will be discussed later, the 
uniform effect in the other cases of the substitution is to depress the 
IsBVorotation, or, in other words, to exert a dextrorotatory, influence. 

3. At the ordinary temperature (20^), the dextrorotatory influence 
of these groups follows in the order 

paratoluyl > benzoyl > metatoluyl > orthotoluyl 
both for methylic and ethylic malate. 

4. At a high temperature (136^), on the other hand, the sequence of 
the dextrorotatory influences of these groups is 

paratoluyl > metatoluyl > benzoyl > orthotoluyl 
in the case of methylic malate, and 

in the case of ethylic malate. 

Thus it will be seen from the diagram on p. 346 that the specific 
rotation curve for methylic benzoylmalate cuts that for methylic meta- 
toluylmalate at a temperature of about 120°, and that the specific 
rotation curve for ethylic benzoylmalate cuts that for ethylic meta- 
toluylmalate at about 95°, and that for ethylic orthotoluylmalate at 
about 132°. Again, it will be seen that the influences of the benzoyl, 
metatoluyl, and orthotoluyl groups become almost identical at the hish 

„.gitized by Google 



Specific notation of Methylie and JEthylic Salts of Benzoyl and Toluyl (o, m, and p) 

Malic Acid, 




za 90. fo. so. 90. 70. 80 00. 100. 110. no. /so. HA 
Temperature in Degrees Centiffrade. 

temperature (136°), both in the case of ethjlic and methylie ma]at€ 

5. At the ordinary temperature (20°), the rotatory influences of 
the several groups are as follows. 


j Methylie malate 
( Ethylic malate 





j Methylie malate 
( Ethylic malate 



( Methylie malate 
( Ethylic malate 



Digitized by Google 



( Methylic malate dextrorotatory 
Paratoluyl } A 

( Ethylic malate ditto. 

Thus in all cases the dextrorotatory influence of the aromatic 
group is greater on the ethylic than on the methylic compound, 
indeed the orthotoluyl group, whilst exerting a dextrorotatory 
influence on ethylic malate, exerts a Isavorotatory influence on 
methylic malate. 

6. At a high temperature (136°), on the other hand, the rotatory 
influences of the several groups are as follows. 






( Methylic malate 

( Ethylic malate 
I Methylic malate 

( Ethylic malate 
I Methylic malate 

( Ethylic malate 
( Methylic malate 









[ Ethylic malate dextrorotatory. 

Thus, again, the dextrorotatory influence of the aromatic group is 
more pronounced on the ethylic than on the methylic compound ; 
indeed, at the high temperature the influence of all four aromatic 
groups is actually IsBvorotatory on methylic malate, whilst the para- 
toluyl group at any rate has still a dextrorotatory influence on 
ethylic malate. 

Mason Ukiyeksitt Collbob, 


XXXIV. — Some Regularities in the Rotatory Power of 

Homologous Series of Optically Active Compounds. 

By Pbecy Fbankland, Ph.D., F.R.S. 

It is difficult to find a consistent explanation for the phenomena to 
which attention has heen drawn in the preceding paper, the lower 
IsBTorotation of the substituted ethylic malates being particularly 
perplexing when it is borne in mind that the leevorotation of methylic 
malate is lower than that of ethylic malate. Thus it would be 
anticipated that the dextrorotatory influence of the aromatic acidyl 
group should depress the levorotation of both methylic and ethylic 
malates, and that the methyl compound should have a smaller Isevo- 

„.gitized by Google 


rotation than the ethyl compound, whilst, as a matter of fact, the 
reverse is actually the case. 

The following explanation is provisionally suggested to account for 
the facts. 

The experimental specific rotations of methylic and ethylic malates 
are not the specific rotations of monomolecular bodies ; both of these 
ethereal salts are associated, and the methyl compound is more highly 
associated than the ethyl compound. This association leads to a lower 
negative specific rotation than is possessed by the monomolecular 
substances, and of the latter the methylic has a higher negative 
rotation than the ethylic. 

On introducing the acid radicles, the molecules become simple or 
more nearly so, and the dextrorotatory influence which they exercise 
depresses the negative rotation of both methylic and ethylic malate, 
the negative rotation of the methylic naturally remaining greater 
than that of the ethylic compound. 

This explanation is also in general harmony with the results 
obtained by Walden {Zeit, physikeU, Chsm., 1895, 17, 245—266) in the 
case of other derivatives of malic acid. Thus, 

Dimethylic malate -C'SS* 

Diethylic „ -lO'lS 

Dipropylic ,, -11'62 

Di-isopropylic,, - 10*41 (about) 

Di-iaobutylic „ -11*14 

Diamylic „ -9"92 

Dicaprylic „ -6*92 (about) 

[«]r [«]r t«]r 

Dimethylic acetylmalate -22*92** propionylmalate -22*94** butyrylmalate -22 •44* 

Diethylic „ -22*52 „ -2220 „ -22*22 

Dipropylic „ -22*86 „ -22*40 

Di-iBobutylic „ -21*88 -21*68 

[«]r [«]r 

Dimethylic isobntyrylmalate -22*36** isovalerylmalate -22*89** 

Diethylic ,, -21*99 „ -22*07 

Dipropylic „ ,, -21*68 

Di-isobutylic „ „ -19*91 

Dimethylic bromacetylmalate -22*40 chlorosnccinate +41*42* 

Diethylic „ -22*48 „ +27*50 

Dipropylic „ -22*24 ,, +25*68 

Di-iflobutylic „ -20*88 „ +21*57 

Diamylic „ — „ +21*66 

* Walden has shown that the chlorosuccinates really corresponding with the 
ordinary malates' hare the same Isyorotations {Ber,, 1896, 29, 138). 

„.gitized by Google 


In the case of all these substituted malates, the substitution raises 
the IffiYorotation, and in each substituted series, excepting that of 
bromacetylmalic acid, the methylic has a higher rotation than the 
ethylic compound. 

The following evidence of the association or otherwise of the 
several compounds referred to in the previous paper is afforded by 
Traube's formula for the calculation of molecular volume.* 

Calculated jlol wt. Experimental 
mol. vol. at ., ,'o,.o mol. vol at 

MethyUc malate 140'6 }^^ 181-2 

„ benzoylnialate 216-6 t4SI^ 218-6 


„ ortho-toluylmalate ... 281-7 _???L 283-6 



meta- „ ... „ TWs ^^'^ 


" P^- ' vim ^^'^ 

* The calculated molecular Yolnmes giTen in this paper have been obtained by 
nsing the formula T^ = m9'9C + nS-lfl" + ^2*80' + g6-60" + 26-9, in which 
(/ = atomic volume of hydrozylic oxygen. The value for (/ becomes 0-4 when OH 
is attached to a carbon atom which is also doubly linked to an atom of o^gen 
(carbozyl group) or when the neighbouring carbon atom is also attached to an OH- 
group, as in the ethereal salts of tartaric acid. (/' = atomic volume of oxygen when 
doubly linked to carbon (either to one or to two carbon atoms). 26*9 = *' Covolume." 
18*2 is deducted for each benzene ring present in the molecule (Ber., 1896, 28, 2724). 

Atomic volume of CI = 13-2 

„ Br = 17-7 {Ber., 29, 1024). 

In my paper in conjunction with Dr. McCrae (Trans., 1898, 73, 324), on some of the 
monacidyl derivatives of ethylic tartrate, the values employed were taken from an 
earlier communication of Traube's, and these led to a wider divergence between the 
calculated and experimental molecular volumes than when the above more recent 
figures are used in calculation. Thus 

Calculated tp^^„-™^«*«i 

^±r^J^ E^ninenUl 

Former. Present. ** ^^'* 

Bthylic monobenzoyltartrate 248*8 262*8 266*6 

„ monopara-toluyltartrate 264-9 268-4 275 '4 

„ monometa' „ „ „ 274*7 

„ monortho- „ „ „ 271*6 

„ dibenzoyltartrate 827-2 842*7 844*6 

„ dipara-toluyltartrate 869*4 874*9 379*6 

„ dimeta- „ „ „ 879*0 

„ diortho- „ „ „ 878*9 

Methylic dibenzoyltartrate 296*0 810*6 307*6 

„ dipara-toluyltartrate 827*2 342*7 841*4 

„ dimetar „ „ „ 341*7 

,, diortho- ,, ,, ,, 886*8 

Methylic tartrate 188*2 189*1 138*6 

Ethylic „ 170*4 171*8 170aOgle 

Vni. l.YYV A A 


Calculated j^^i ^|. EzperimentAl 

moL voL at j^^oimq mol. vol. at 

15«. »15 /* 16'. 

Ethylic malate 172-8 t4?L- l^^'S 

•^ 1-1890 

I, benxoylmakte....! 247*8 ^^- 252-8 

H ortho-toluylmalate ... 268-9 -~^ 269 2 

II meta* „ /*. „ ^^ 269*4 

.par*. ., . ... „ ^M. 269-8 

From these figures, it appears that both ethylic and methylic malate 
are associated, and the latter much more strongly than the former, 
whilst the aromatic acidylmalates are not associated, their experi- 
mental molecular volumes being, indeed, in excess of the calculated 
values. This being the case, it is not improbable that the lavorotation 
of unassociaUd methylic malate may be greater than the lievo.* 
rotation of unassociated ethylic malate, and, in fact, that the real 
values of [ajo for the monomolecular ethereal salts of malic acid 
diminish in passing up the series, from the methylic term. This view 
is borne out by applying Traube's formula to the known members of 
the series as prepi^ed and examined by Walden {loc. cit.). It should 
be noted that Walden's densities were all determined at 20^, so that 
the experimental molecular volumes for that temperature are a little* 
higher than they would be for 15^, but the difference is too small to 
render recalculation worth while at this stage of the inquiry. 

Calculated j^^]^ ^ Experimental 
mol. vol* at. -Q^o/.o' mol. toI. at 

iMmethyUc malate 140*6 ^~ 181-8 

Diethylio „ 172-8 -jM- 168-2 

Dipwpylic , 206-0 ^^ 202-9 

Di-iwpiopylic malate 205-0 ^^^^ 202-6 

Di-iflobutyUc , 287-2 ^^ 286-1 

l>iamylic ., 269-4 ^mW) ^^^*'^^* 

Dicaprylic „ 8660 ^^ 866-8 

The above figures show that, in ascending the series, the terms 
exhibit less and less evidence of association; indeed, the experi- 

* Then can be little doubt that this figure is incorrect, through the density (1 *079) 
being wrongly given in Walden'a paper. The density should obviously lie between 
those of the butylic and caprylic compounds. /^ ^ ^ ^T ^ 

Digitized by VjO(J*vIC 


mental molecular volume of dicaprylic malate is somewhat in 
excess of the calculated. 

The ethereal salts of malic acid are generally cited as an instance 
of an homologous series exhibiting a 'maximum rotation' (Guye 
and Chavanne, Btdl. Soc. Chim., 1896, [iii], 15, 190), and, as is 
well known, these maxima occurring in homologous series have been 
used by Guye to support his views concerning the "prodtust of 
cuymmetry" In the present instance, however, the maximum 
exhibited in the series of ethereal salts of malic acid is explicable on 
the very probable assumption, based on molecular volume, of associa- 
tion occurring in a diminishing degree from the methylic compound 
upwards. Moreover, the product of asymmetry does not predict a 
maximum rotation within this series, but, on the contrary, demands 
that the rotations sdiould continuously diminish in ascending the 
series, thus, 

Product of asymmetry. 

Dimethylic malate —193 

DiethyHc „ -171 

Dipropylic „ —148 

DibutyUc „ -128 

Diamylic „ —Ill 

Dicaprylic „ —75 

From the following tables, it will be seen that, according to 
Traube's formula, the substituted malates described by Walden 
are either not associated at all, or only slightly so, even in the case 
of the methylic compound. 

Ethereal salts of acetylmalic 

ctcid. (Walden, loc. cU*) 



Product of 

MoL vol. 


for 15". 

Mol. vol. 
eip. for 20*. 














Ethereal salts of propionylmalic acid, (Walden, loe. cU.) 

Dimethylic . 
lylic .... 












Product of 

Mol. voL 


for 16^ 

MoL YoL 
exp. for 20*. 

Ethereai aaba qf hUyrylmalic <ieid. (Wfdden, loc. cU,) 



Dipropylic .... 
Di-iflobatylic , 













Eiheredl scUts qf isobuiyrylmalie ctcid. (Wfdden, loe. di.) 

DiethyUc .. 






Ethereal salia qf isovaieryhnalio acid. (Walden, loc. cit,) 



Dipropylic... . 
Di-isobatylic . 


_ 22-89'* 











Ethereal salts qf broTnacetybnaltc add. (WaldeOy loe. eit.) 

Dimethylic .., 




















Ethereal saits of ehlorosuccinio acid, (Wfdden, loc. cU.) 



Di-isoDutylic. . 


-h 27*60 
+ 21*66 




The phenomena in the case of the chlorosuocinateB would appear to 
admit of twofold interpretation : either the normal rotation of di- 
methylic chlorosuocinate is so much in excess of that of the ethylic 
compound that the association of the methyl compound does not 
depress its rotation below that of the ethylic chlorosuocinate, or the 
association actually leads in this case to an increase of dextrorotation. 

The two series of the ethereal salts of methoxy- and ethoxy-suodnio 
acids may also be regarded as substituted malates and considered 
along with the preceding ones ; for these, the following data baye 
been obtained by Purdie and Williamson (Trans., 1895, ffiT, 971)^ 

_.„„y Google 



Product of 



Mol. Tol. 


for 16^ 

Mol. ToL 

Ethereal salts qf mUhoocysuccinie cid. 




Bntylic (nonn.} 




Bntylic (norm.) 

+ 52-6r(atl2') 
+ 6a-ll (at 18") 
+ 46-21 (at 16") 
+ 41-68 (atl6") 



^§8=168-1 (for 12-) 




r0149 = 2W-2(f<>'10 

Ethereal salts of ethoxysueeinic acid. 

+ 69-86'* (at 18") 
+ 66-29 (at 17") 
+ 61-81 (at 16") 
+46-48 (at 6") 



-:j555 = 171-9(forl8") 




* The density is only given for 6° (1 -0046), the value for 16" has heen calculated 
on the assumption that the density diminishes hy 0*0008 for each 1" rise in 

These two series are particularly interesting; they each exhibit 
continuously diminishing rotations, whilst the product of asymmetry 
predicts a maximum at the propylic term for the first, and a maximum 
for a higher term in the case of the second series. In each series, the 
methylic term alone furnishes much evidence of association, but in 
neither series does this association lead to the methylic compound 
having a lower rotation than the ethylic. However, in each series the 
rotation of the methylic compound is anomalous, thus the difference 
in rotation between the methylic and ethylic methoxysuccinate is less 
than between the other consecutive terms of the series, whilst the 
difference in the case of the methylio and ethylic ethoxysuccinate is 
greater than that between the next two consecutive terms of this series 
Moreover, the methylic compound with the most anomalous rotation 
has also the most anomalous molecular volume. 

Thus this very justifiable assumption with regard to the association 
and consequent masked rotation of the initial terms of the series of 
unsubstituted malates is in general harmony with the fact that in the 

„.gitized by Google 


several series of substituted malates the methylic compound has 
almost without exception the highest IsBvorotation. 

The question next arises as to whether those homologous series in 
which the product of asymmetry demands a maximum rotation, and 
in which a maximum actually occurs, also exhibit this phenomenon of 
association in the initial terms. 

Of series exhibiting such a maximum rotation approximately in 
harmony with the demands of the product of asymmetry, one of the 
first to be prepared and most fully investigated is that of the ethereal 
salts of glyceric acid, thus (P. Franklatid and MacGregor, Trans, 1893, 
63, 511 and HIO ; P. Frankland and Price, Trans, 1897, 71, 253). 

Eiheredl salts of glyceric add. 





Butyhc (norm.)... 


Amylic (secondary 

butylmethyl) .. 
Heptylic (norm.) 







- 12-94 




Product of 



+ 288-8 
+ 844-8 
+ 858-2 
+ 846-8 
+ 846-8 


+ 241-8 

Mol. TOl. 


for U\ 

102 1 


MoL Tol. 

at 15** 




Thus all the known terms of this series exhibit association, the 
latter being most pronounced, as before, in the case of the lower 
members, and becoming less and less marked as the series is ascended. 
If it be assumed that this association depresses the rotation, it 
obviously becomes highly doubtful whether the maximum rotation, 
which the series exhibits, has any real connection with the maximum 
indicated by the products of asymmetry. 

Again, another homologous series which has been very fully investi- 
gated is that of the ethereal salts of diacetylglycerio acid ; it exhibits 
a very marked maximum rotation, but at a term very far removed 
from that predicted by the product of asymmetry. The following (next 
page) are the values which have been found (P. Frankland and 
MacGregor, Trans., 1894, 65, 750 ; P. Frankland and Price, loo, eU.). 

This series is peculiarly instructive, as it shows how unconnected is 
the phenomenon of maximum rotation with the maximum product of 
asymmetry. Thus, whilst the maximum rotation falls on the isobutyl 
term (as in the series of glycerates), the maximum product of asym- 
metry is not reached until the term C^gH^^. There is also much less 

„.gitized by Google 

Ethereal 9alt8 qf dMoetylglyemic add. 






Amylic (seoondaiy 

Oc^uc (nonn.) ... 










-16 '87 

Flrodnct of 


P X 10«. 

+ 17-4 



+ 126*2 

Mol. ToL 


for 16^ 

M oL Tol. 

at 16* 






evidence of association in the case of the methylio compoand than in 
the case of the glycerates, and coincidentallj with this the rotations 
of the initial terms of the diacetylglycerio series are less markedly 
inferior to the maximum rotation than in the glyceric series. 



Prodnot of 

Mol. Tol. 


for 16'. 

Mol. TOl. 

Ethereal ealte qf dimonoiMoraceiylglyeerie add. (P. Frankland and 
Patterson, Trans., 1898, 73» 181.) 

Butylic . 

1*4268 (16') 
1*8698 (16') 

-16*80 (16') 






209-6 (16') 

Bihertal ealts qf di-dichloracetylglyceric add, (P. Frankland and 
Patterson, loe. dt.) 

Batylio . 

1*6290 (16') 

- 18*96' ao 

-18*88 (16*8') 

+ 86 
+ 28 
+ 10 



Ethereal ealis qf dirtriehloracetylglycerto add. (P. Frankland and 
Patterson, loe, di.) 


1*6118 (12*) 
1*6602 (12') 

-14-17' (12') 
-18*66 (12-6") 

+ 69 
+ 68 
+ 88 


266*0 (12') 
274 1 (12') 

Digitized by VjOOQIC 


In the above three series, the relations are essentially similar to those 
found in the diacetylglycerates ; there is only evidence of association 
in the case of the dimonochloraoetylglycerates, and in all cases 
the methylic compound hacr a distinctly lower rotation than the 
ethylic. The Isevorotation of the methylic and ethylic glycerates is, 
in fact, simply augmented by a certain number of degrees in each case. 

It will be noticed that the molecular volumes calculated by Traube's 
formula generally fall very far short of the experimental vidues in the 
case of these chlorine compounds. This abnormally large molecular 
volume of compounds containing much chlorine has already been 
pointed out by Traube. 

Another very similar series is that of the ethereal salts of di- 
benzoylglyceric acid, of which several terms have been prepared and 
studied (P. Frankland and MacGregor, Trans., 1896, 69, 104; 
P. Frankland and Price, Trans., loc. eii,). 

Ethereal salte qf dibenzaylglyeeric 




Product of 



MoL vol. 


for 16°. 

Mol. vol. 

at 16° 





+ 26-89° 
+ 21-00 


+ 61-8° 
+ 37-4 
+ 19-1 

+ 1-2 








Amylic (secondary 


There is thus evidence of association throughout the known terms 
of the series. In this series, the product of asymmetry reaches a 
minimum at the hezylic term, after which it again increases, the 
rotations diminish from the methylic term upwards, but the rotation 
of the methylic term is distinctly anomalous, the difference between 
it and the ethylic being much smaller than that between the ethylic 
and propylic terms. Presumably, the normal rotation of all the 
members of this series is higher than that given above, the rotation of 
the methylic compound being the greatest ; indeed, it has been shown 
(P. Frankland and Pickard, 1896, 69, 123) that in glacial acetic acid 
(in which fairly normal molecular weights were obtained by the 
cryoscopic method) the rotation of methylic dibenzoylglycerate 
ranged from [a]^* +32-38'' (<5 = i8'6 per cent.) to +-34'34° (c = l-7. 
per cent.), whilst in benzene solution (in which the molecular weights 
were distinctly below the theoretical) the rotation ranged from 
[a]o= + 40-72° (c = 34-l per cent.) to + 45-70° (c = 3-0 per cent.). 

Digitized by VjOOQIC 

Ethereal aaUa qf lactie add. (Walker, Trans., 1895, 67, 916.) 





1-100 (18°/4°) 
1-004 (19°/4°) 


-17 06 

Prodnct of 


Mol. vol. 


for 15°. 

MoL Yol. 


94-6 (18°) 
114-6 (19") 

Ethereal salts of acetyllactie acid. (P. Frankland and Henderson, 

Proc., 1895, 54.) 

EthyKc* ... 

1-0627 (16°/16°) 




138-2 (16°) 
162-0 (16°) 

Ethereal salts of chloropropionie acid. (Walker, Trans., loc. eit.) 


1^144 (1674°) 
1-0821 (16°/4°) 
1-0524 (16°/4°) 

al salts qf brai 

1-816 (14°/4°) 

+ 26-88° 
+ 19-88 
+ 11-00 


: +42-66° 

1 +81-46 

+ 21-98 


le acid. fW 

+ 278 
+ 90 


''alker, loe 

115 7 

107-1 (16°) 


126-1 (16°) 


148-0 (16°) 

112*7 (17°) 




180-6 (19«») 


148-8 (14°) 

* The lactic acid from which these compounds were prepared was afterwards 
found to be partially racemised, so that the absolute values given for the rotations 
ue, of coarse, too small, but the ratio between the two is probably correct. 

N.6. — It has been shown by Purdie and Williamson (Trans., 1896, 
09, 837) that the change of sign in the case of the chloro- and 
bromo-propionates is due to inversion, as in the case of the chloro - 
succinates from malic acid. 

The relations exhibited by the series of lactates and acetyllactates 
are very similar to those already drawn attention to in the case of 
the malates and acidylmalates. Thus the rotation of methylic lactate 
is less than that of ethylic lactate, whilst methylic acetyllactate has 
a higher rotation than ethylic acetyllactate. The lower rotation of 
methylic than of ethylic lactate, it will be observed, is coincident with 
a much greater degree of association of the methylic than of the 
ethylic compound, whilst the amount of association in the case of the 
acetyllactates is not nearly so great. It may, therefore, be inferred 

Digitized by VjOOQIC 


that in the ficetyllaotates the association of the methylic compound 
is not sufficient to depress its rotation below that of the ethylic 
compound. The phenomena in the case of the chloro- and bromo- 
propionates are similar to those exhibited by the acetyllactates. 

There are two homologous series of active compounds to which great 
prominence has been given by Guye as supporting the connection 
between maximum product of asymmetry and maximum rotation; 
these are the ethereal salts of active amylic alcohol and of active 
valeric acid respectively. The data given in the two following tables 
have been furnished by the investigations of Guye and Chavanne 
{Bull. Soc. Ghim., 1896, [iii], 16, 177, 273), 

Ethereal salts of active amylic alcohol. (Guye and Chavanne 

, loe. oU.) 



Product of 

MoL vol. 


for 15^ 

Mol. voL 


c alcohol 





106-8 {IV) 
181-6 (20') 


+ 2 01 





+ 2-58 



149-0 (20*) 
166-6 (20T 

propionate ... 

+ 2-77 



n-bntyrate ... 

0-862 (20*) 

+ 2-69 





0-860 (20**) 

+ 2-52 



200-0 (20*) 


0-869 (20**) 

+ 2-40 



216-6 (20': 1 

?i-heptylate ... 

0-861 (20**) 

+ 2-21 


280 1 

232-3 120* 


0-860 (20**) 

+ 2-10 



248-8 1 20* 


0-861 (20*) 

+ 1-96 



264-8 (20«» 


0-871 (20") 

+ 1-88 


293-9 (20" 

T^dnodecylate . 

0-859 (20*') 

+ 1-56 



314-8 (20°:i 

n-palmitate ... 

0-864 (20^ 

+ 1-28 



881-7 (20^ 
414-0 (20") 


0-866 (20°) 

+ 1-27 



In this series, therefore, the product of asymmetry attains a 
maximum at the acetate term, whilst the maximum dextrorotation 
is actually exhibited by the following, or propionate, term. On 
comparing the calculated and experimental molecular volumes, it will 
be seen that the first few terms are tainted with the evidence of 
association, and it is highly probable that the rotations which they 
exhibit are not those of the single molecules. Thinking that possibly 
the l»vorotation of amylic alcohol itself might be due to association, 
and that the single molecule was perhaps dextrorotatory like the 
other terms of the series, I have had its rotation determined in 
alcoholic solution by Mr. Mason, B.Sc., with the result that a more 
highly laavorotatory figure was obtained. As the amylic alcohol 
would presumably be less associated in alcoholic solution than in the 

Digitized by VjOOQIC 


pure state, it may be inferred that monomolecular amylic alcohol has 
a higher lievorotation than [a]D» — 4'62.* In this case, then, the 
above series starts with a lievorotation, which becomes diminished 
by successive additions to the variable group until the propionate 
term of the series is reached, beyond which further additions to the 
variable group slowly increase the lasvorotation, or, rather, diminish 
the dextrorotation. Regarded in this light, the series is Isevorotatory, 
its ascent being attended by diminution of the Invorotation, a 
minimum laBVorotation being reached at the propionate term. This 
is, as far as I am aware, the only series in which a minimum has 
actually been discovered (see, in this connection, p. 369). 

Ethereal scUts qf (ictive valeric acid. (Guye and Chavanne.) 



Product of 

Mol. vol. 


for 16". 

Mol. vol. 

Valeric acid 

0-938 (22*) 

+ 18-64* 



108-7 (22") 

(in liquid 


+ 17-3" 

(in aqneons 


Methylio valerate ... 

0-882 (22") 





0-864 (22") 

+ 18-44 



160-6 (22") 

Propylic „ 

0-860 (22") 

+ 11-68 



167-4 (22") 

0-866 (22") 

+ 10-60 



184-6 (22") 

Isobntylic „ 


+ 10-48 

. 861 


184-8 (22") 

Benzylic „ 

0-982 (22') 

+ 5-31 



196-6 (22°) 

This series exhibits very similar relations to those discussed above 
for the ethereal salts of active amylic alcohol. The product of 
asymmetry attains a maximum for ethylio valerate, whilst the maximum 
rotation is actually found in the case of methylic valerate, that is, for 
the term just preceding that which the product of asymmetry predicts. 
It will, however, be observed that the rotation of methylic is very 
much in excess of that of ethyiic valerate, so that, apparen1;}y, the 
slight association, of which there is evidence in the case of methylic 
valerate, is not sufficient to absolutely depress its rotation below that 
of ethyiic valerate. In the case of liquid valeric acid itself, however, 
there is evidence of much greater association than for methylic 
valerate, and probably this degree of association is sufficient to depress 

♦ It has, further, been quite recently shown by Guye and Emily Aston (Abstr., 
1898, ii, 469) that amylic alcohol in aqueous solution gives a value [o]d = -6*1", 
whilst in benzene solution it is [a]i> = -4-6, thus corroborating the above assump- 
tion as to th^ higher lievorotation of unassociated amylio alcohol. 

„.gitized by Google 


the rotation of yaleric acid below that of the methylic oompound. 
This interpretation is, to some extent, borne out by the fact that 
aqueous yaleric acid has actually a higher rotation than metiiylic 
valerate, although the rotation may in this case be complicated by the 
circumstance that, in aqueous solution, the valeric acid is partially 

Another series cited by M. Guye (he. eit.) as furnishing a verifica- 
tion of the predictions of the product of asymmetry is that of the 
active amy lie oxides, the data for which are contained in the following 

Active amylie oxidei. (Guye and Ohavanne, loe, eii.) 




Product of 

Mol. vol. 


for 16'. 

Mol. ToL 


0-764 (18*) 
0-769 (18°) 
0-788 (18°) 
0-798 (22°) 
0-778 (22°) 
0-806 (22°) 
0-911 (22°) 

+ 0-89' 
+ 0-61 
+ 0-90 
+ 1'SS 
+ 0-96 
+ 0-81 
+ 1-82 



186-8 (18*) 
162-8 (18T 
166-0 (18*) 
180-4 (22°) 
186-8 (22*) 
196-4 (22') 







M^V*M*,J «•«» J « 

This series, it will be seen, has a maximum product of asymmetry 
in the case of the propylamyl term, whilst the maximum rotation 
was found for the next higher one, namely, normal butylamyl oxide. 
The series presents certain features which do not occur in any 
of the others to which I have previously referred, for whilst the 
methyl and ethyl terms exhibit no evidence of association, such 
evidence is exhibited in a slight degree by the propyl, and in a higher 
degree by the normal butyl, term, or, in other words, precisely for 
that term where the maximum rotation occurs. The suspicion is 
naturally aroused that this maximum rotation is connected with the 
abnormal molecular volume of the term in question. It will be 
noticed, also, that the isobutyl compound, which bears no evidence of 
associi^ion, has a much lower rotation than the normal butyl com- 
pound. The fact of the maximum rotation, however, remains, for the 
rotation even of the isobutyl compound is greater than that of any of 
the other terms, excepting the benzyl, which, in consequence of its 
very different structure, may be excluded from consideration for the 
present. There is, however, another circumstance which must be 
taken into account before accepting this series as a verification of the 
predictions furnished by the product of asymmetry. From Guye and 
Chavanne's paper, it appears that the methyl, ethyl, propyl, and cetyl 
compounds were all prepared by the action of sodium amylate on the 

„.gitized by Google 


iodides of the respective radicles, whilst the normal and isobutyl 
compounds were obtained by the action of amylic bromide on sodium 
butylate and isobutylate respectively. Thus, in the preparation of all 
those oxides which yielded the lowest rotations, sodium was made to 
act on amylic alcohol^ a procedure which is well known to be the most 
effective method of rendering this alcohol inactive, whilst in the 
preparation of the two butyl compounds, which gave the highest 
rotations, the amylic alcohol was not submitted to the direct action of 
sodium. There can be no doubt that this circumstance renders it 
necessary to receive the rotation results obtained with reserve. 

In the following tables, the values for the several known series of 
ethereal salts derivable from tartaric acid are investigated ; the data 
are those given by Freundler (Thesis, Paris, 1894). 



Product of 


P X 10«. 

Mol. YoL 


for 16°. 

Mol. Yol. 

Ethereal scUta of tartoHc €md. 




Butylic ..... 

1-3284 (20*) 
1-098 (19°) 

+ 2-14(20°) 

+ 7-66(20°) 

+ 12-44(20°) 

+ 14-88(20°) 

+ 10-8 (19°) 


139-1 184-0(20°) 

171-3 170-8(20°) 
203-6 206-8 (20°) 
207-1 (20°) 
286-7 238-6 (19° 

Ethylic ... 
Propylic ... 
Bnlyiic ... 

Ethylic .. 
Propylic .. 
Bat^lio .. 

Ethylic . 
Propylic ., 
Bn^lic r 

-10-7° (16°) 

+ 118 

+ 0-4 (16°) 

+ 6-6 (16°) 


+ 6-9 (12°) 


+ 11-4 (19°) 


Ethereal salts of dipropionyltartaric acid. 

1-181 (16°) 





Ethereal salts of dibutyryltartaric acid. 





841 1 



-16-1° (13°) 

+ 194 


' -0-8 (16°) 

+ 92 

1-067 (16-6°) 

+ 5-2 (18°) 


+ 6-0 (14°) 



+ 8-5 (19°) 

246-5 (16°) 
282-9 (14°) 
3161 (15°) 
860-2 (15-6°) 
848-5 (16-5°) 

813-1 (15-6°) 
860-5 i;i6-6°) 
883-6 (16°) 
382-8 (16°) 

Ethereal salts qf divaleryltartaric add. 

1-101 (18°) 
1-081 (18°) 
1-082 (18-6°) 

+ 8-3 

+ 4-8 
+ 7-4 


+ 248 

+ 168 

+ 78 

308-9 814-2 (18°) 

341-1 850-2(12°) 

873-3 882-8 (165°) 

406-5 417-1 (18°) 

„ 416-7 (18-6°) 

Digitized by VjOOQIC 




Product of 


P X 10«. 

Mol. vol. 


for 15'. 

Mol. vol. 

Ethyfic .... 
Propylio .... 
Isobntylic . 

JStliereal salts qf dicaproyltartaric acid. 

1-040 (14-6°) 

-15-9** (16") 
-3-1 (16") 
+ 2-2 (16") 
+ 6-0 (16") 

+ 282 
+ 216 

+ U1 
+ 7 


846-9 (14") 
418-7 (16") 
462-1 (13") 

Ethereal acUta qf di^sovaleryltariaric acid. 

Ethylic .. 
Propylio .. 

Ethylic .. 
Propylic .. 

1-028 (18-0") 





+ 0-7 


+ 6-7 


+ 248 

+ 168 

+ 78 



Bihereal scdts of di-^obwtyryltartoHc €kcid» 





+ 2-2 


+ 8-4 


+ 194 

+ 92 



312-6 (16-6") 
860-6 (17-6") 
418-8 (18-0") 

850-8 (16-0") 
888-6 (16-6") 

* The methylic compound being solid, its rotation was only determined in 
alcoholic solution. 



Product of 

Mol. voL 


for 16' 

Mol. YoL 

Ethereal salts qf di-monochloracetyltaHarie add, (P. Frankland and 
Patterson, Trans., 1898, 73, 193.) 

Ethylic ... 
Propylic ... 
Bulylic ... 


-0 80" ao 

+ 7-40 (16-6") 

+ 221 

+ 128 

+ 88 



232-8 (14") 

Ethereal salts of di-dichJoracetyliartaric acid, (P. Frankland and 
Patterson, loc, cit., 190.) 

Ethylic .., 
Propylic .,, 
Butylic .. 

1-6108 (16") 

+ 11-97" (19*2") 
+ 16-80 (16") 

+ 301 
+ 261 
+ 187 
+ 120 


264-7 (16") 
302-7 (16") 

Digitized by VjOOQIC 




Product of 

Mol. vol. 


for 16°. 

Mol. vol. 

Ethereal salts of moTKhtrichlaraoetf/ltartaric acid, 
Patterson, loc cit., 186.) 

(P. Frankland and 

Kthylic ... 
Propylic ... 
Bu^lic ... 


+ 8-29*(17°) 
+ 16-80 (12°) 

-206— 880 =-686 

-124— 420 =-644 

-67— 440 =-607 

-32— 441 =-478 


214-5 (17°) 
252-4 (16°) 

In the case of the neutral tartrates and diacidyltartrates, the 
gioupe attached to each of the two asymmetric carbon atoms 
being identical, the product of asymmetry given in the above tables 
refers to one carbon atom only, whilst in the monacidyltartrates, the 
groups attached to the two asymmetric carbon atoms being different, 
each of these carbon atoms has a product of asymmetry of its own. 
Thus, this is' the case with the ethereal salts of monotrichloracetyl- 
tartaric acid abov^. 

The above table shows that, in the simple tartrates, there is evidence 
of association in the case of the methylio compound, and in all the 
other series of derivatives the molecular volume of the methylic 
compound, although in excess of that calculated by Traube's formula, 
is markedly less in excess than are the molecular volumes of the 
highw homolgues. 

In the case of methylic tartate, again, there is some independent 
evidence that association leads to a movement of the rotatory power 
in the negative direction, for it has been shown by Freundler (loc. 
cU.) that methylic tartrate is very highly associated in benzene 
solution (a molecular weight of 411 instead of the theoretical 178 
having been obtained) and that in benzene solution, also, the specific 
rotation is — 6'P instead of +2'14% the value found for the pure 
substance in the fused state. This inference is based, of course, on 
the assumption that methylic tartrate is more associated in benzene 
solution than in the fused state, an assumption which is justifiable 
if Traube's formula be capable of revealing the degree of association. 

That a substance may be more associated in solution than in the 
liquid state appears not to be without precedent, thus the molecular 
weight of phenol as determined by the ftamsay and Shields method is 
given by Eamsay and Aston (Trans., 1894, 66, 168) as 

MoL wt. 

46 — 78° 
78 —131-7 


133-6 ) 
126-9 l 
110-9 J 




whilst the following values were ohtained hy Beckmann {Zeit. 
pkysikal. Chem.^ 1888, 2, 727 — 730) with the cryoscopic method in 
benzene solution (melting point of benzene = 5*44). 

in benzene eolation, 


of freezing point. 


Hoi. wt 



















On reviewing the figures given in the above tables, it is impossible 
not to be impressed with the fact that it is only in the case of those 
series in which the product of asymmetry attains a maximum at or 
about the propylic or butylic terms that there is any even ap- 
proximate coincidence between this maximum and the maximum 
rotation in the series, and the suspicion naturally arises as to whether 
this maximum rotation is not really dependent on the associated 
nature of the first few terms of the series leading to abnormal rota- 
tions of these initial terms, and not to the existence of a maximum 
product of asymmetry at all. 

This suspicion is increased by the circumstance that in series, like 
that of the tartrates, in which the product of asymmetry continuously 
diminishes from the first term, there is, notwithstanding, a maximum 
rotation at or about the same region (that is, the propylic) of the 
series. Again, in the series of the diacetylglycerates in which the 
product of asymmetry does not attain a maximum until the thirteenth 
term, the maximum rotation is already exhibited by the fourth or 
butylic term. 

It would appear, therefore, that the coincidence between theory and 
experiment, in the matter of maximum product of asymmetry and 
maximum rotation, may be quite accidental, and the idea that Guye's 
suggestive speculation receives any support from this occasional 
coincidence must at present be received with much reserve. 

If the conception of the product of asymmetry be laid aside, it would 
be anticipated that, in ascending an homologous series, the rotation 
should either continuously increase or decrease, the increments and 
decrements respectively becoming less and less marked with growing 
molecular weight. On examining the several series referred to in the 
tables above, it will be seen that this is actually the case in the 

1. — ^Acetylmalates (diminishing rotations, excepting a slight rise in 
the case of the propylic term, series known up ta^sobu^lic). „ 


2. — Propionylmalates (diminishing rotations, but only the methylic 
and ethylic terms known). 

3. — Bntyrylmalates (diminishing rotations, excepting a slight rise in 
the case of the propylic term, series known ap to isobutylic). 

4.— Isobutyrylmalates (diminishing rotations, but only the methylic 
and ethylic terms known). 

6. — Isovalerylmalates (diminishing rotations, series known up to 
isobntylio term). 

6. — Bromacetylmalates (diminishing rotations, excepting a slight 
rise in the case of the ethylic term, series known up to isobutylic). 

7. — Ghlorosuccinates (diminishing rotations, series known up to 
diamylic term). 

8. — Methoxysuocinates (diminishing rotations, series known up to 
normal butylic term). 

9. — Ethoxysuccinates (diminishing rotations, series known up to 
normal butylic term). 

10. — ^Benzoylmalates (diminishing rotations, but only the methylic 
and ethylic terms known). 

11. — Orthotoluylmalates (diminishing rotations, but only methylic 
and ethylic terms knowji). 

12. — Metatoluylmalates (diminishing rotations, but only methylic and 
ethylic terms known). 

13. — Paratoluylmalates (diminishing rotations, but only methylic 
and ethylic terms known). 

14. — Dibenzoylglycerates (diminishing rotations, known up to 
amylic term). 

15. — Acetyllactates (diminishing rotations, but only methylic and 
ethylic terms known). 

16. — Chloropropionates (diminishing rotations, series known up to 
propylic terms). 

17. — Bromopropionates (diminishing rotations, series known up to 
propylic term). 

18. — Valerates, excluding valeric acid itself (diminishing rotations, 
series known up to isobutylic term). 

19. — Dimonochloracetylglycerates (increasing rotations, but only 
methylic and ethylic terms known). 

20. — Di-dichloracetylglycerates (increasing rotations, but only 
methylic and ethylic terms known. 

21. — Di-trichloracetylglycerates (increasing rotations, but only 
methylic and ethylic terms known). 

22. — ^Di-dichloracetyltartrates (increasing rotations, but only methylic 
and ethylic terms known). 

23. — Mono-trichloraoetyltartrates (increasing rotations, but only 
methylic and ethylic terms known). 

VOL. hXHyi. C^r^c^n]o 

Digitized by VjO(J*V IC 


24. — Dibenzoyltartrates (diminishing rotations, series known up to 
bntylic term). 

25. — Di-orthotoluyltartrates (diminishing rotations, only methylic 
and ethjlic terms known). 

26. — Di-metatoluyltartrates (diminishing rotations, only methylic 
and ethylic terms known). 

27. — ^Di-paratoluyltartrates (diminishing rotations, only methylic 
and ethylic terms known). 

In ascending some series, although the influence on rotation is 
continuously in the same direction, the process is attended with a 
change of sign. 

1. — Dipropionyltartrates (series known up to butylic term^ change 
of sign at ethylic term). 

2. — Dibutyryltartrates (series known up to butylic term, change of 
sign at propylic term). 

3. — Divaleryltartrates (series known up to butylic term, change of 
sign at propylic term). 

4. — ^Dicaproyltartrates (series known up to butylic term, change of 
sign at propylic term). 

5. — Di-isovaleryltartrates (series known up to butylic term, change 
of sign at propylic term). 

6. — Di-isobutyryltartrates (series known up to butylic term, change 
of sign at propylic term). 

7. — Di-monochloracetyltartrates (only methylic and ethylic terms 
known, change of sign at ethylic term). 

In ascending other series, again, a maximum rotation is attained 
which is followed by decline. 

1. — Malates. The maximum is reasonably explicable as due to the 
association of the initial terms. 

2. — Lactates. The maximum is reasonably explicable as due to 
the association of the initial terms. 

3. — ^Tartrates. The maximum, although coincident with associa- 
tion of the early terms, is too pronounced to be thus accounted for. 

4. — Glycerates. The maximum, although coincident with associa- 
tion of the early terms, is too pronounced to be thus accounted for. 

5. — Diacetylglyceratee. The maximum, although coincident with 
association of the early terms, is too pronounced to be thus accounted 

6. — Amylic oxides. Botation results doubtful in consequence of 
variable mode of preparation. 

The grounds for asserting above that the maximum rotation 
exhibited by the malates and lactates respectively is ''reasonably 
explicable as due to the association of the initial terms " are (1) that 
strong evidence, of such association is furnished by the molecular 

Digitized by VjOOQ IC 


▼olamesy and (2) that in the substitution compounds (acetylmalates, 
alkozystioeinatesy acetyllactates, ohloropropionates, Ac.) in which there 
is little or no evidence of association, the rotation of the methylic is 
greater than that of the ethylic compound, whilst the rotations of 
methylic mlQate and lactate are inferior to those of ethylic malate 
and lactate respectively. In fact, in the case of the malates, 
some substitutions raise the lievorotation whilst others depress 
ity but in both cases the resulting rotations point to methylic 
malate having a greater Isovorotation than ethylic malate. The latter 
statement is at first sight negatived by the fact that the chlorosucci- 
natesy alkozysuccinates, chloropropionates, and bromopropionates 
have a dextrorotation diminishing from the methylic term upwards, 
which would appear to mean that these substitutions had shifted the 
rotation in the positive direction by a certain amount in each case, 
and that, starting with the lower rotation of methylic malate and 
lactate respectively, this had, therefore, resulted in a higher dextro- 
rotation for methylic chloro^uccinate, chloropropionate, and bromo- 
propionate than for the corresponding ethylic compounds. We know, 
however, that in the transformation of the malates into the chloro- 
Buocinates, and of the lactates into the ohloro- and bromo-propionates, 
an tfUMmon takes place, and that the ohlorosuocinates, chloro- and 
bromo-propionates really corresponding with the original malates and 
lactates from which they are derived are more highly IsBVorotatory 
than those malates and lactates themselves. In the case of the 
alkoxysuocinates, it is not known with certainty to which of the 
optically isomeric malates they are directly related, but I venture to 
suggest from their analogy to the chloroeuccinates that the dextro- 
rotatory alkoxysuocinates correspond with the dextromalates, and 
the recent investigations of Purdie and Lander (Trans., 1898, 73, 
287) in which alkoxysuocinates of high levorotation were obtained 
by acting with alkyl iodides on ordinary (Isevo-) silver malate leave 
little room for doubt that this is actually the case. 

On the other hand, I have stated above that the maximum rotation 
exhibited in the series of tartrates, glycerates, and diacetylglycerates 
cannot satisfactorily be referred to the occurrence of association in the 
ease of the earlier terms of these series, because (1) the maximum is 
so very pionounced, and (2) because in the numerous substitution com- 
pounds which are known, and which exhibit little or no evidence of 
association, the relationship between the rotation of the methylic and 
ethylic compounds is retained. Thus methylic glycerate has a lower 
Isvorotation than ethyHc glycerate. When the levorotation is 
increased, as by the introduction of two acetyl, chloracetyl, or dichlor- 
aoetyl groups, the methylic compound remains throughout of inferior 

B B 2 

Digitized by VjOOQIC 


lavorotation to that of the ethylic compound. Again, when the l»vo- 
rotation is depressed (op rather exchanged for a strong dextro- 
rotation), as by the introduction of two benzoyl groups, the methylic 
compound retains its lower laBVorotation, that is, has a higher dextro- 
rotation than the ethylic compound. 

Similarly, methylic tartrate has a lower dextrorotation than ethylic 
tartrate ; and when the dextrorotation is increased, as by the intro- 
duction of two dichloracetyl groups, or one trichloracetyl group, the 
dextrorotation of the methylic compound remains inferior to that of 
the ethylic compound, whilst when the dextrorotation is depressed 
(or even exchanged for laevorotation), as in the majority of the 
diacidyltartratea, the dextrorotation of the methylic compound 
remains inferior to that of the ethylic, or, in other words, in these 
cases the methylic compound has a higher Iffivorotation than the 

But whilst this occurrence of a maximum cannot thus be explained 
away in the case of the glycerates, diacetylglycerates, and tartrates, it 
can also not be justifiably used as a confirmation of the predictions of 
the product of asymmetry, for in the series of tartrates the product of 
asymmetry diminishes in ascending from the methylic term, in the 
diacetylglycerates the maximum product of asymmetry is not reached 
until the thirteenth term, whilst in the glycerates alone is there 
approximate coincidence between maximum product of asymmetry and 
maximum rotation. Is it not equally, or more, possible that the 
maximum rotation in these series is dependent on the number of 
atoms attached in a chain to the asymmetric carbon atom f According 
to the commonly accepted views of stereochemistry, a continuous chain 
of five carbon atoms will all but return upon itself, and, beyond this 
further additions to the chain will lead to such interference as must 
necessitate a readjustment of the exact positions occupied by the 
carbon atoms in a shorter chain. It is surely highly probable that 
this stereochemical change should be betrayed by some irregularity in 
the rotatory manifestations, for example, by the exhibition of a maxi- 
mum rotation in those series in which the ascent of the series leads to 
an increase in the rotatory power. ^ 

It is obvious, too, that in those series the ascent of which is 
attended with continuous diminution in the rotation, some irregu- 
larity — a minimum rotation — is to be anticipated on stereochemical 
grounds when the growing chain attached to the asymmetric carbon 
atom attains the magnitude of five atoms or thereabouts. So far, 
however, the number of known terms in the several series of this 
kind is insufficient to detect this irregularity or minimum rotation. 
The series of the ethereal salts of amylic alcohol, as pointed out oi\ 

Digitized by VjOOQIC 


p. 358y and that of the diacet jltartrates * may be regarded as exhibit- 
ing a minimum rotation, for growth of the one group leads to 
diminution of the original rotation of the first member of the series, 
hot this diminution leads at once to the zero line being crossed, and 
with snccessiye additions to the group there is increasing rotation 
of opposite sign (equivalent to diminishing rotation of the original 
sign) until a maximum rotation of the opposite sign (equivalent to 
minimum rotation of the original sign) is attained, on which follows 
diminishing rotation of the opposite sign {equivalent to increasing 
rotation of the original eign). 

The rotation phenomena exhibited in the ascent of homologous 
series would appear to be capable of classification into the following 
three types. 

a. — Increase in rotatory power up to a maximum, followed by a 

decline becoming more or less asymptotic, as indicated by the 

curves in Fig. 1. 

Fio. i. 

To this type belong the glycerates, diacetylglycerates, and tartrates ; 
these series are actually known beyond the occurrence of the maxi- 

* The data concarning the rotation of the diacetyltartrates aie somewhat im- 
perfect, and they have, in consequence, not been included in the preceding tables. 
The following figores are given by Frenndler {loe, eit,). 

Meihylic diacdyUaHraU in solution in absolute alcohol t=26°, 1=2, c=2*4818, 
«i>=-46', [al>=-15-r. 

&h/ylic diaeeiffUartraU in superfused state ^=25% 1=1, ao= +5% whilst in solu- 
tion In absolute alcohol i=26% 1=2, c=7*088, a»= -f-8', [«]»= +0*4^ 

Prvpylie diaeetyltartrate in superfused state ai>=-i-18'4, 1=1, whilst in solution 
in absolute aleohol t=2S\ 1=2, 0=4-9872, a»= -1-57', [a]i>= -H9'6°. 

BiUyliediaeetyUartr(U6in}ixiTddstBAet=ie^,l=l,d=l'09e,ao= +8-8°,[a]p= +S 0% 
whilst in solution in absolute alcohol t=22\ 1=2, e-6'9425, ao=+l'* 12', 

laapropylie diaeetyltartrtUe in solution in absolute alcohol t=20% 1=2, e=6*5885, 
ap=+4r, [al,=:-h6-9°. 

Isobuiylic diaedyUa/rWaU in liquid state <=18'6% 1=2, (2=1'096, a»= +87'' 8', 
[al!= +17-0% in solution in absolute alcohol t=l^, 1=2, c=s:6'4669, Op= +r 18', 

These figures show inoontestably that, for the normal series of radicles, the maxi- 
mum dextrorotation is possessed by the propylio term. 

Digitized by VjOOQIC 


mum. It is to be anticipated that, when sof&ciently explored, the 

dimonochloracetylglyoerates, di-dichloracetylglyoerates, di-trichlor- 

acetjlglycerates, di-dichloracetyltartrates, and monotrichloraoetyltar- 

trates will each exhibit a maximum in the vicinity of the bntylic 


b. — ^Diminution in rotatory power down to a minimum, followed by 

a rise with diminishing increments, as indicated by the curves in 

Fig. 2. 

Fio. 2. 

To this type belong the chlorosuccinates, ethoxysuccinates, 
methoxysuccinates, chloropropionates, bromopropionates, acetyllae- 
tates, acidylmalates, benzoylglycerates, and valerates. Thus in 
ascending these series, as far as they are known, the rotations 
continuously diminish, but it is to be anticipated that they will 
respectively exhibit a minimum when they are more fully explored. 

Fio. 8. 

c. — ^Diminution in rotatory power, accompanied sooner or later by 
change of sign, the series exhibiting a maximum with a sign the 
reverse of that of the initial member of the series, as indicated by 
the curves in Fig. 3. 

To this type belong the diacetyltartrates, which series b^ins with 
a strong l»vorotation for the methylic term, the ethylic term being 
already dextrorotatory, and this dextrorotation attains a maximum 

Digitized by 



for the propylio term, normal butylic diacetyltartrate being already 
of distinctly inferior dextrorotation. 

Ohange of sign is similarly exhibited by the dipropionyltartrates 
dibtttyryltartratesy dioaproyltartrates, di-isobatyryltartrates, and di- 
isovaleryltartrateSy also by the dimonoohloracetyltartrates, but in 
none of these has the maximum dextrorotation yet been disp